SlideShare a Scribd company logo
Sungkyunkwan University
Department of Human ICT Convergence
Yoon Sup Choi, Ph.D.
How to implement the digital medicine in the future
: measure, collect and interpret patient-generated data
The Convergence of IT, BT and Medicine
How to Implement the Digital Medicine in the Future
Inevitable Tsunami of Change
http://rockhealth.com/2015/01/digital-health-funding-tops-4-1b-2014-year-review/
http://rockhealth.com/2015/07/2-1b-digital-health-funding-first-half-2015-keeping-pace-2014/
What is most important factor in digital medicine?
“Data! Data! Data!” he cried.“I can’t
make bricks without clay!”
- Sherlock Holmes,“The Adventure of the Copper Beeches”
Three Steps to Implement Digital Medicine
• Step 1. Measure the Data
• Step 2. Collect the Data
• Step 3. Insight from the Data
Step 1. Measure the Data
Smartphone: the origin of healthcare innovation
2013?
The election of Pope Benedict
The Election of Pope Francis
The Election of Pope Francis
The Election of Pope Benedict
SummerTanThese Days
How to Implement the Digital Medicine in the Future
Jan 2015 WSJ
CellScope’s iPhone-enabled otoscope
PEEK (portable eye examination kit)
http://www.peekvision.org
OScan: oral cancer detection
How to Implement the Digital Medicine in the Future
Kinsa Smart Thermometer
SpiroSmart: spirometer using iPhone
iPhone Breathalyzer
AliveCor Heart Monitor
Sleep Cycle
BeyondVerbal: Reading emotions from voices
How to Implement the Digital Medicine in the Future
Fitbit
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
iRythm ZIO patch
How to Implement the Digital Medicine in the Future
Google’s Smart Contact Lens
C8 Medisensor: non-invasive blood glucose sensor
Withings Wireless Blood Pressure Monitor
Huinno: Cuff-less Blood Pressure Monitor
Smart Band detecting seizure
n
n-
ng
n
es
h-
n
ne
ne
ct
d
n-
at
s-
or
e,
ts
n
a-
gs
d
ch
Nat Biotech 2015
Personal Genome Analysis
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
Results within 6-8 weeksA little spit is all it takes!
DTC Genetic TestingDirect-To-Consumer
120 Disease Risk
21 Drug Response
49 Carrier Status
57Traits
$99
Health Risks
Health Risks
Health Risks
Drug Response
Traits
음주 후 얼굴이 붉어지는가
쓴 맛을 감지할 수 있나
귀지 유형
눈 색깔
곱슬머리 여부
유당 분해 능력
말라리아 저항성
대머리가 될 가능성
근육 퍼포먼스
혈액형
노로바이러스 저항성
HIV 저항성
흡연 중독 가능성
Ancestry Composition
Neanderthal Ancestry
23andMe Customer Growth
http://goldbio.blogspot.kr/2014/12/pg-100.html
How to Implement the Digital Medicine in the Future
Step1. Measure the Data
• With your smartphone
• With wearable devices (connected to smartphone)
• Personal genome analysis
... without even going to the hospital!
Step 2. Collect the Data
How to Implement the Digital Medicine in the Future
Sci Transl Med 2015
How to Implement the Digital Medicine in the Future
Google Fit
Samsung SAMI
How to Implement the Digital Medicine in the Future
Epic MyChart App Epic EHRDatabaseDexcom App
Withings App
Dexcom CGM
Nike+
Patients Device/Apps
HealthKit EHR Hospital
Whitings
+
• Data stored in DB on the iPhone (, not mirroring to the cloud)
• Consumer controls what data goes in/out, privacy level
• HealthKit connects/direct devices, store data based on privacy rules
Apple Watch
iPhone
How to Implement the Digital Medicine in the Future
Sci Transl Med 2015
How to Implement the Digital Medicine in the Future
Without cloud computing,
we cannot collect the real-time big data from the patients
The Regulations
Practice Fusion, an EMR based on the cloud
Step 3. Insight from the Data
Data Overload
How to Analyze and Interpret the Big Data?
and/or
Two ways to get insights from the big data
Hospitals in the future: Data Analysis Center
Doctors in the future: Data Scientists
No choice but to bring AI into the medicine
How to Implement the Digital Medicine in the Future
AliveCor Heart Monitor
“AliveCor has received an additional FDA
510(k) clearance, this time for an algorithm
that allows its smartphone ECG to detect
atrial fibrillation with high accuracy.”
“the algorithm has a 100 percent sensitivity
(it never returns a false negative) and a 97
percent specificity (it returns false positives
about 3 percent of the time). For obvious
reasons, the algorithm was designed to err
on the side of false positives”
DeepFace: Closing the Gap to Human-Level
Performance in FaceVerification
Taigman,Y. et al. (2014). DeepFace: Closing the Gap to Human-Level Performance in FaceVerification, CVPR’14.
Figure 2. Outline of the DeepFace architecture. A front-end of a single convolution-pooling-convolution filtering on the rectified input, followed by three
locally-connected layers and two fully-connected layers. Colors illustrate feature maps produced at each layer. The net includes more than 120 million
parameters, where more than 95% come from the local and fully connected layers.
very few parameters. These layers merely expand the input
into a set of simple local features.
The subsequent layers (L4, L5 and L6) are instead lo-
cally connected [13, 16], like a convolutional layer they ap-
ply a filter bank, but every location in the feature map learns
a different set of filters. Since different regions of an aligned
image have different local statistics, the spatial stationarity
The goal of training is to maximize the probability of
the correct class (face id). We achieve this by minimiz-
ing the cross-entropy loss for each training sample. If k
is the index of the true label for a given input, the loss is:
L = log pk. The loss is minimized over the parameters
by computing the gradient of L w.r.t. the parameters and
by updating the parameters using stochastic gradient de-
Human: 95% vs. DeepFace in Facebook: 97.35%
Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)
FaceNet:A Unified Embedding for Face
Recognition and Clustering
Schroff, F. et al. (2015). FaceNet:A Unified Embedding for Face Recognition and Clustering
Human: 95% vs. FaceNet of Google: 99.63%
Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)
False accept
False reject
s. This shows all pairs of images that were
on LFW. Only eight of the 13 errors shown
the other four are mislabeled in LFW.
on Youtube Faces DB
ge similarity of all pairs of the first one
our face detector detects in each video.
False accept
False reject
Figure 6. LFW errors. This shows all pairs of images that were
incorrectly classified on LFW. Only eight of the 13 errors shown
here are actual errors the other four are mislabeled in LFW.
5.7. Performance on Youtube Faces DB
We use the average similarity of all pairs of the first one
hundred frames that our face detector detects in each video.
This gives us a classification accuracy of 95.12%±0.39.
Using the first one thousand frames results in 95.18%.
Compared to [17] 91.4% who also evaluate one hundred
frames per video we reduce the error rate by almost half.
DeepId2+ [15] achieved 93.2% and our method reduces this
error by 30%, comparable to our improvement on LFW.
5.8. Face Clustering
Our compact embedding lends itself to be used in order
to cluster a users personal photos into groups of people with
the same identity. The constraints in assignment imposed
by clustering faces, compared to the pure verification task,
lead to truly amazing results. Figure 7 shows one cluster in
a users personal photo collection, generated using agglom-
erative clustering. It is a clear showcase of the incredible
invariance to occlusion, lighting, pose and even age.
Figure 7. Face Clustering. Shown is an exemplar cluster for one
user. All these images in the users personal photo collection were
clustered together.
6. Summary
We provide a method to directly learn an embedding into
an Euclidean space for face verification. This sets it apart
from other methods [15, 17] who use the CNN bottleneck
layer, or require additional post-processing such as concate-
nation of multiple models and PCA, as well as SVM clas-
sification. Our end-to-end training both simplifies the setup
and shows that directly optimizing a loss relevant to the task
at hand improves performance.
Another strength of our model is that it only requires
False accept
False reject
Figure 6. LFW errors. This shows all pairs of images that were
incorrectly classified on LFW. Only eight of the 13 errors shown
here are actual errors the other four are mislabeled in LFW.
5.7. Performance on Youtube Faces DB
We use the average similarity of all pairs of the first one
hundred frames that our face detector detects in each video.
This gives us a classification accuracy of 95.12%±0.39.
Using the first one thousand frames results in 95.18%.
Compared to [17] 91.4% who also evaluate one hundred
frames per video we reduce the error rate by almost half.
DeepId2+ [15] achieved 93.2% and our method reduces this
error by 30%, comparable to our improvement on LFW.
5.8. Face Clustering
Our compact embedding lends itself to be used in order
to cluster a users personal photos into groups of people with
the same identity. The constraints in assignment imposed
by clustering faces, compared to the pure verification task,
Figure 7. Face Clustering. Shown is an exemplar cluster for one
user. All these images in the users personal photo collection were
clustered together.
6. Summary
We provide a method to directly learn an embedding into
an Euclidean space for face verification. This sets it apart
from other methods [15, 17] who use the CNN bottleneck
layer, or require additional post-processing such as concate-
nation of multiple models and PCA, as well as SVM clas-
Constructing higher-level
contextual/relational features:
Relationships between epithelial
nuclear neighbors
Relationships between morphologically
regular and irregular nuclei
Relationships between epithelial
and stromal objects
Relationships between epithelial
nuclei and cytoplasm
Characteristics of
stromal nuclei
and stromal matrix
Characteristics of
epithelial nuclei and
epithelial cytoplasm
Building an epithelial/stromal classifier:
Epithelial vs.stroma
classifier
Epithelial vs.stroma
classifier
B
Basic image processing and feature construction:
H&E image Image broken into superpixels Nuclei identified within
each superpixel
A
Relationships of contiguous epithelial
regions with underlying nuclear objects
Learning an image-based model to predict survival
Processed images from patients Processed images from patients
C
D
onNovember17,2011stm.sciencemag.orgwnloadedfrom
TMAs contain 0.6-mm-diameter cores (median
of two cores per case) that represent only a small
sample of the full tumor. We acquired data from
two separate and independent cohorts: Nether-
lands Cancer Institute (NKI; 248 patients) and
Vancouver General Hospital (VGH; 328 patients).
Unlike previous work in cancer morphom-
etry (18–21), our image analysis pipeline was
not limited to a predefined set of morphometric
features selected by pathologists. Rather, C-Path
measures an extensive, quantitative feature set
from the breast cancer epithelium and the stro-
ma (Fig. 1). Our image processing system first
performed an automated, hierarchical scene seg-
mentation that generated thousands of measure-
ments, including both standard morphometric
descriptors of image objects and higher-level
contextual, relational, and global image features.
The pipeline consisted of three stages (Fig. 1, A
to C, and tables S8 and S9). First, we used a set of
processing steps to separate the tissue from the
background, partition the image into small regions
of coherent appearance known as superpixels,
find nuclei within the superpixels, and construct
Constructing higher-level
contextual/relational features:
Relationships between epithelial
nuclear neighbors
Relationships between morphologically
regular and irregular nuclei
Relationships between epithelial
and stromal objects
Relationships between epithelial
nuclei and cytoplasm
Characteristics of
stromal nuclei
and stromal matrix
Characteristics of
epithelial nuclei and
epithelial cytoplasm
Epithelial vs.stroma
classifier
Epithelial vs.stroma
classifier
Relationships of contiguous epithelial
regions with underlying nuclear objects
Learning an image-based model to predict survival
Processed images from patients
alive at 5 years
Processed images from patients
deceased at 5 years
L1-regularized
logisticregression
modelbuilding
5YS predictive model
Unlabeled images
Time
P(survival)
C
D
Identification of novel prognostically
important morphologic features
basic cellular morphologic properties (epithelial reg-
ular nuclei = red; epithelial atypical nuclei = pale blue;
epithelial cytoplasm = purple; stromal matrix = green;
stromal round nuclei = dark green; stromal spindled
nuclei = teal blue; unclassified regions = dark gray;
spindled nuclei in unclassified regions = yellow; round
nuclei in unclassified regions = gray; background =
white). (Left panel) After the classification of each
image object, a rich feature set is constructed. (D)
Learning an image-based model to predict survival.
Processed images from patients alive at 5 years after
surgery and from patients deceased at 5 years after
surgery were used to construct an image-based prog-
nostic model. After construction of the model, it was
applied to a test set of breast cancer images (not
used in model building) to classify patients as high
or low risk of death by 5 years.
www.ScienceTranslationalMedicine.org 9 November 2011 Vol 3 Issue 108 108ra113 2
onNovember17,2011stm.sciencemag.orgDownloadedfrom
Digital Pathologist
Sci Transl Med. 2011 Nov 9;3(108):108ra113
Digital Pathologist
Sci Transl Med. 2011 Nov 9;3(108):108ra113
Top stromal features associated with survival.
primarily characterizing epithelial nuclear characteristics, such as
size, color, and texture (21, 36). In contrast, after initial filtering of im-
ages to ensure high-quality TMA images and training of the C-Path
models using expert-derived image annotations (epithelium and
stroma labels to build the epithelial-stromal classifier and survival
time and survival status to build the prognostic model), our image
analysis system is automated with no manual steps, which greatly in-
creases its scalability. Additionally, in contrast to previous approaches,
our system measures thousands of morphologic descriptors of diverse
identification of prognostic features whose significance was not pre-
viously recognized.
Using our system, we built an image-based prognostic model on
the NKI data set and showed that in this patient cohort the model
was a strong predictor of survival and provided significant additional
prognostic information to clinical, molecular, and pathological prog-
nostic factors in a multivariate model. We also demonstrated that the
image-based prognostic model, built using the NKI data set, is a strong
prognostic factor on another, independent data set with very different
SD of the ratio of the pixel intensity SD to the mean intensity
for pixels within a ring of the center of epithelial nuclei
A
The sum of the number of unclassified objects
SD of the maximum blue pixel value for atypical epithelial nuclei
Maximum distance between atypical epithelial nuclei
B
C
D
Maximum value of the minimum green pixel intensity value in
epithelial contiguous regions
Minimum elliptic fit of epithelial contiguous regions
SD of distance between epithelial cytoplasmic and nuclear objects
Average border between epithelial cytoplasmic objects
E
F
G
H
Fig. 5. Top epithelial features. The eight panels in the figure (A to H) each
shows one of the top-ranking epithelial features from the bootstrap anal-
ysis. Left panels, improved prognosis; right panels, worse prognosis. (A) SD
of the (SD of intensity/mean intensity) for pixels within a ring of the center
of epithelial nuclei. Left, relatively consistent nuclear intensity pattern (low
score); right, great nuclear intensity diversity (high score). (B) Sum of the
number of unclassified objects. Red, epithelial regions; green, stromal re-
gions; no overlaid color, unclassified region. Left, few unclassified objects
(low score); right, higher number of unclassified objects (high score). (C) SD
of the maximum blue pixel value for atypical epithelial nuclei. Left, high
score; right, low score. (D) Maximum distance between atypical epithe-
lial nuclei. Left, high score; right, low score. (Insets) Red, atypical epithelial
nuclei; black, typical epithelial nuclei. (E) Minimum elliptic fit of epithelial
contiguous regions. Left, high score; right, low score. (F) SD of distance
between epithelial cytoplasmic and nuclear objects. Left, high score; right,
low score. (G) Average border between epithelial cytoplasmic objects. Left,
high score; right, low score. (H) Maximum value of the minimum green
pixel intensity value in epithelial contiguous regions. Left, low score indi-
cating black pixels within epithelial region; right, higher score indicating
presence of epithelial regions lacking black pixels.
onNovember17,2011stm.sciencemag.orgDownloadedfrom
and stromal matrix throughout the image, with thin cords of epithe-
lial cells infiltrating through stroma across the image, so that each
stromal matrix region borders a relatively constant proportion of ep-
ithelial and stromal regions. The stromal feature with the second
largest coefficient (Fig. 4B) was the sum of the minimum green in-
tensity value of stromal-contiguous regions. This feature received a
value of zero when stromal regions contained dark pixels (such as
inflammatory nuclei). The feature received a positive value when
stromal objects were devoid of dark pixels. This feature provided in-
formation about the relationship between stromal cellular composi-
tion and prognosis and suggested that the presence of inflammatory
cells in the stroma is associated with poor prognosis, a finding con-
sistent with previous observations (32). The third most significant
stromal feature (Fig. 4C) was a measure of the relative border between
spindled stromal nuclei to round stromal nuclei, with an increased rel-
ative border of spindled stromal nuclei to round stromal nuclei asso-
ciated with worse overall survival. Although the biological underpinning
of this morphologic feature is currently not known, this analysis sug-
gested that spatial relationships between different populations of stro-
mal cell types are associated with breast cancer progression.
Reproducibility of C-Path 5YS model predictions on
samples with multiple TMA cores
For the C-Path 5YS model (which was trained on the full NKI data
set), we assessed the intrapatient agreement of model predictions when
predictions were made separately on each image contributed by pa-
tients in the VGH data set. For the 190 VGH patients who contributed
two images with complete image data, the binary predictions (high
or low risk) on the individual images agreed with each other for 69%
(131 of 190) of the cases and agreed with the prediction on the aver-
aged data for 84% (319 of 380) of the images. Using the continuous
prediction score (which ranged from 0 to 100), the median of the ab-
solute difference in prediction score among the patients with replicate
images was 5%, and the Spearman correlation among replicates was
0.27 (P = 0.0002) (fig. S3). This degree of intrapatient agreement is
only moderate, and these findings suggest significant intrapatient tumor
heterogeneity, which is a cardinal feature of breast carcinomas (33–35).
Qualitative visual inspection of images receiving discordant scores
suggested that intrapatient variability in both the epithelial and the
stromal components is likely to contribute to discordant scores for
the individual images. These differences appeared to relate both to
the proportions of the epithelium and stroma and to the appearance
of the epithelium and stroma. Last, we sought to analyze whether sur-
vival predictions were more accurate on the VGH cases that contributed
multiple cores compared to the cases that contributed only a single
core. This analysis showed that the C-Path 5YS model showed signif-
icantly improved prognostic prediction accuracy on the VGH cases
for which we had multiple images compared to the cases that con-
tributed only a single image (Fig. 7). Together, these findings show
a significant degree of intrapatient variability and indicate that increased
tumor sampling is associated with improved model performance.
DISCUSSION
Heat map of stromal matrix
objects mean abs.diff
to neighbors
H&E image separated
into epithelial and
stromal objects
A
B
C
Worse
prognosis
Improved
prognosis
Improved
prognosis
Improved
prognosis
Worse
prognosis
Worse
prognosis
Fig. 4. Top stromal features associated with survival. (A) Variability in ab-
solute difference in intensity between stromal matrix regions and neigh-
bors. Top panel, high score (24.1); bottom panel, low score (10.5). (Insets)
Top panel, high score; bottom panel; low score. Right panels, stromal matrix
objects colored blue (low), green (medium), or white (high) according to
each object’s absolute difference in intensity to neighbors. (B) Presence
R E S E A R C H A R T I C L E
onNovember17,2011stm.sciencemag.orgDownloadedfrom
Top epithelial features.The eight panels in the figure (A to H) each
shows one of the top-ranking epithelial features from the bootstrap
anal- ysis. Left panels, improved prognosis; right panels, worse prognosis.
‘Minority Report (2002)’
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
Data Baby
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
+
Integration of Health Data and Genomic Data
+ +
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
• Apple HealthKit
• Fitbit Data
• Personal Genome Data
• GPS
+
Personalized
Healthcare
Advices
How to Implement the Digital Medicine in the Future
How to Implement the Digital Medicine in the Future
Three Steps to Implement Digital Medicine
• Step 1. Measure the Data
• Step 2. Collect the Data
• Step 3. Insight from the Data
How to Implement the Digital Medicine in the Future
Feedback/Questions
• Email: yoonsup.choi@gmail.com
• Blog: http://www.yoonsupchoi.com
• Facebook: Yoon Sup Choi
How to Implement the Digital Medicine in the Future

More Related Content

What's hot

Electronic Health Record Standardisation in India
Electronic Health Record Standardisation in IndiaElectronic Health Record Standardisation in India
Electronic Health Record Standardisation in India
Apollo Hospitals Group and ATNF
 
Electronic health record powerpoint assignment for informatics
Electronic health record powerpoint assignment for informaticsElectronic health record powerpoint assignment for informatics
Electronic health record powerpoint assignment for informatics
Michaelina Alexander
 
Remote Patient Monitoring (RPM) - Enabling New Models of Care
Remote Patient Monitoring (RPM) - Enabling New Models of Care Remote Patient Monitoring (RPM) - Enabling New Models of Care
Remote Patient Monitoring (RPM) - Enabling New Models of Care
Anthony Fanning
 
Remote patient monitoring technology
Remote patient monitoring technologyRemote patient monitoring technology
Remote patient monitoring technology
remote healthtech
 
Clinical information system-final copy
Clinical information system-final copyClinical information system-final copy
Clinical information system-final copy
CISgroup
 
Accelerating Patient Care with Real World Evidence
Accelerating Patient Care with Real World EvidenceAccelerating Patient Care with Real World Evidence
Accelerating Patient Care with Real World Evidence
CitiusTech
 
Remote patient monitoring in home health
Remote patient monitoring in home healthRemote patient monitoring in home health
Remote patient monitoring in home health
Samantha Haas
 
QUALITY INDICATOR IN NURSING.pptx
QUALITY INDICATOR IN NURSING.pptxQUALITY INDICATOR IN NURSING.pptx
QUALITY INDICATOR IN NURSING.pptx
anjalatchi
 
Evaluation of A Clinical Information System
Evaluation of A Clinical Information SystemEvaluation of A Clinical Information System
Evaluation of A Clinical Information System
nrodrock
 
CLINICAL INFORMATICS ppt
CLINICAL INFORMATICS pptCLINICAL INFORMATICS ppt
CLINICAL INFORMATICS ppt
Bhavitha Pulaparthi
 
Introduction to Telemedicine
Introduction to TelemedicineIntroduction to Telemedicine
Introduction to Telemedicine
Devang Parikh
 
Digital Health Care Technology
Digital Health Care TechnologyDigital Health Care Technology
Digital Health Care Technology
Nawanan Theera-Ampornpunt
 
Digital health
Digital healthDigital health
Digital health
Dr.Puvaneswari kanagaraj
 
Remote patient monitoring system
Remote patient monitoring system Remote patient monitoring system
Remote patient monitoring system
Rahul Singh
 
Enterprise Master Patient Index - IBM White Paper
Enterprise Master Patient Index - IBM White PaperEnterprise Master Patient Index - IBM White Paper
Enterprise Master Patient Index - IBM White Paper
Bart de Witte
 
Telemedicine
TelemedicineTelemedicine
Telemedicine
Darshil Shah
 
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEMLABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
Aj Raj
 
The Future of Digital Health in 2022
The Future of Digital Health in 2022The Future of Digital Health in 2022
The Future of Digital Health in 2022
Diana Girnita
 
Telemedicine
TelemedicineTelemedicine
Telemedicine
ADITYA .
 
Telemedicine ppt
Telemedicine pptTelemedicine ppt
Telemedicine ppt
khandhar
 

What's hot (20)

Electronic Health Record Standardisation in India
Electronic Health Record Standardisation in IndiaElectronic Health Record Standardisation in India
Electronic Health Record Standardisation in India
 
Electronic health record powerpoint assignment for informatics
Electronic health record powerpoint assignment for informaticsElectronic health record powerpoint assignment for informatics
Electronic health record powerpoint assignment for informatics
 
Remote Patient Monitoring (RPM) - Enabling New Models of Care
Remote Patient Monitoring (RPM) - Enabling New Models of Care Remote Patient Monitoring (RPM) - Enabling New Models of Care
Remote Patient Monitoring (RPM) - Enabling New Models of Care
 
Remote patient monitoring technology
Remote patient monitoring technologyRemote patient monitoring technology
Remote patient monitoring technology
 
Clinical information system-final copy
Clinical information system-final copyClinical information system-final copy
Clinical information system-final copy
 
Accelerating Patient Care with Real World Evidence
Accelerating Patient Care with Real World EvidenceAccelerating Patient Care with Real World Evidence
Accelerating Patient Care with Real World Evidence
 
Remote patient monitoring in home health
Remote patient monitoring in home healthRemote patient monitoring in home health
Remote patient monitoring in home health
 
QUALITY INDICATOR IN NURSING.pptx
QUALITY INDICATOR IN NURSING.pptxQUALITY INDICATOR IN NURSING.pptx
QUALITY INDICATOR IN NURSING.pptx
 
Evaluation of A Clinical Information System
Evaluation of A Clinical Information SystemEvaluation of A Clinical Information System
Evaluation of A Clinical Information System
 
CLINICAL INFORMATICS ppt
CLINICAL INFORMATICS pptCLINICAL INFORMATICS ppt
CLINICAL INFORMATICS ppt
 
Introduction to Telemedicine
Introduction to TelemedicineIntroduction to Telemedicine
Introduction to Telemedicine
 
Digital Health Care Technology
Digital Health Care TechnologyDigital Health Care Technology
Digital Health Care Technology
 
Digital health
Digital healthDigital health
Digital health
 
Remote patient monitoring system
Remote patient monitoring system Remote patient monitoring system
Remote patient monitoring system
 
Enterprise Master Patient Index - IBM White Paper
Enterprise Master Patient Index - IBM White PaperEnterprise Master Patient Index - IBM White Paper
Enterprise Master Patient Index - IBM White Paper
 
Telemedicine
TelemedicineTelemedicine
Telemedicine
 
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEMLABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
LABORATORY INFORMATION SYSTEM RADIOLOGY INFORMATION SYSTEM
 
The Future of Digital Health in 2022
The Future of Digital Health in 2022The Future of Digital Health in 2022
The Future of Digital Health in 2022
 
Telemedicine
TelemedicineTelemedicine
Telemedicine
 
Telemedicine ppt
Telemedicine pptTelemedicine ppt
Telemedicine ppt
 

Viewers also liked

1.cue.pandora.part1.v2.0 final
1.cue.pandora.part1.v2.0 final1.cue.pandora.part1.v2.0 final
1.cue.pandora.part1.v2.0 final
Jim "Brodie" Brazell
 
LIBER and EU projects
LIBER and EU projectsLIBER and EU projects
LIBER and EU projects
LIBER Europe
 
Quality Service 2017
Quality Service 2017Quality Service 2017
Quality Service 2017
Quality Services
 
C# Programming with Visual Studio 2005
C# Programming with Visual Studio 2005C# Programming with Visual Studio 2005
C# Programming with Visual Studio 2005
LearnItFirst.com
 
Primera carrera ucv 10 y 5km-Copa Direccion Deporte
Primera carrera ucv 10 y 5km-Copa Direccion DeportePrimera carrera ucv 10 y 5km-Copa Direccion Deporte
Primera carrera ucv 10 y 5km-Copa Direccion Deporte
CarreraycaminataUCV
 
OMNITRACKER GIS Gateway
OMNITRACKER GIS GatewayOMNITRACKER GIS Gateway
OMNITRACKER GIS Gateway
OMNINET USA
 
Nº1 kumua magazine digital ING
Nº1 kumua magazine digital INGNº1 kumua magazine digital ING
Nº1 kumua magazine digital ING
Alberto Lopez
 
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
Whyte's
 
Brochure ats
Brochure atsBrochure ats
140212 Sesión inicial asignatura de packaging
140212 Sesión inicial asignatura de packaging140212 Sesión inicial asignatura de packaging
140212 Sesión inicial asignatura de packaging
Fernando Monzon
 
Repartido n°5 la civilización egipcia
Repartido n°5 la civilización egipciaRepartido n°5 la civilización egipcia
Repartido n°5 la civilización egipcia
Fernando de los Ángeles
 
Remote IP Power Switches
Remote IP Power SwitchesRemote IP Power Switches
Remote IP Power Switches
Chris Barber
 
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
Kienpin Nelly Hung Sam
 
Tivoly Plaza
Tivoly PlazaTivoly Plaza
Tivoly Plaza
romeroviedo
 
그렇게 나는 스스로 기업이 되었다
그렇게 나는 스스로 기업이 되었다그렇게 나는 스스로 기업이 되었다
그렇게 나는 스스로 기업이 되었다
Yoon Sup Choi
 
Presentación Esfera BPO Solutions
Presentación Esfera BPO SolutionsPresentación Esfera BPO Solutions
Presentación Esfera BPO Solutions
nicocaste
 
Etat des lieux des énergies renouvelables dans les nouvelles régions
Etat des lieux des énergies renouvelables dans les nouvelles régionsEtat des lieux des énergies renouvelables dans les nouvelles régions
Etat des lieux des énergies renouvelables dans les nouvelles régions
PEXE
 
Es difícil ser un buen profesor
Es difícil ser un buen profesorEs difícil ser un buen profesor
Es difícil ser un buen profesor
José Zamora Pérez
 
Open sistemas es_v2011
Open sistemas es_v2011Open sistemas es_v2011
Open sistemas es_v2011
Alvaro Garcia
 
La Estrategia Del Oceano Azul[2]
La Estrategia Del Oceano Azul[2]La Estrategia Del Oceano Azul[2]
La Estrategia Del Oceano Azul[2]
Progresista
 

Viewers also liked (20)

1.cue.pandora.part1.v2.0 final
1.cue.pandora.part1.v2.0 final1.cue.pandora.part1.v2.0 final
1.cue.pandora.part1.v2.0 final
 
LIBER and EU projects
LIBER and EU projectsLIBER and EU projects
LIBER and EU projects
 
Quality Service 2017
Quality Service 2017Quality Service 2017
Quality Service 2017
 
C# Programming with Visual Studio 2005
C# Programming with Visual Studio 2005C# Programming with Visual Studio 2005
C# Programming with Visual Studio 2005
 
Primera carrera ucv 10 y 5km-Copa Direccion Deporte
Primera carrera ucv 10 y 5km-Copa Direccion DeportePrimera carrera ucv 10 y 5km-Copa Direccion Deporte
Primera carrera ucv 10 y 5km-Copa Direccion Deporte
 
OMNITRACKER GIS Gateway
OMNITRACKER GIS GatewayOMNITRACKER GIS Gateway
OMNITRACKER GIS Gateway
 
Nº1 kumua magazine digital ING
Nº1 kumua magazine digital INGNº1 kumua magazine digital ING
Nº1 kumua magazine digital ING
 
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
Whyte's Fine Art Auctioneers Dublin 20 September 2014 THE ECLECTIC COLLECTOR ...
 
Brochure ats
Brochure atsBrochure ats
Brochure ats
 
140212 Sesión inicial asignatura de packaging
140212 Sesión inicial asignatura de packaging140212 Sesión inicial asignatura de packaging
140212 Sesión inicial asignatura de packaging
 
Repartido n°5 la civilización egipcia
Repartido n°5 la civilización egipciaRepartido n°5 la civilización egipcia
Repartido n°5 la civilización egipcia
 
Remote IP Power Switches
Remote IP Power SwitchesRemote IP Power Switches
Remote IP Power Switches
 
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
Eduweb 2012 - Los metaversos como recursos pedagógicos en la asesoróa persona...
 
Tivoly Plaza
Tivoly PlazaTivoly Plaza
Tivoly Plaza
 
그렇게 나는 스스로 기업이 되었다
그렇게 나는 스스로 기업이 되었다그렇게 나는 스스로 기업이 되었다
그렇게 나는 스스로 기업이 되었다
 
Presentación Esfera BPO Solutions
Presentación Esfera BPO SolutionsPresentación Esfera BPO Solutions
Presentación Esfera BPO Solutions
 
Etat des lieux des énergies renouvelables dans les nouvelles régions
Etat des lieux des énergies renouvelables dans les nouvelles régionsEtat des lieux des énergies renouvelables dans les nouvelles régions
Etat des lieux des énergies renouvelables dans les nouvelles régions
 
Es difícil ser un buen profesor
Es difícil ser un buen profesorEs difícil ser un buen profesor
Es difícil ser un buen profesor
 
Open sistemas es_v2011
Open sistemas es_v2011Open sistemas es_v2011
Open sistemas es_v2011
 
La Estrategia Del Oceano Azul[2]
La Estrategia Del Oceano Azul[2]La Estrategia Del Oceano Azul[2]
La Estrategia Del Oceano Azul[2]
 

Similar to How to Implement the Digital Medicine in the Future

Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
Digital Future of the Surgery: Brining the Innovation of Digital Technology i...Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
Yoon Sup Choi
 
Prediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep LearningPrediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep Learning
IRJET Journal
 
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
IRJET Journal
 
Research Inventy : International Journal of Engineering and Science
Research Inventy : International Journal of Engineering and ScienceResearch Inventy : International Journal of Engineering and Science
Research Inventy : International Journal of Engineering and Science
inventy
 
IRJET- Breast Cancer Prediction using Deep Learning
IRJET-  	  Breast Cancer Prediction using Deep LearningIRJET-  	  Breast Cancer Prediction using Deep Learning
IRJET- Breast Cancer Prediction using Deep Learning
IRJET Journal
 
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
sadique_ghitm
 
Iaetsd multi-view and multi band face recognition
Iaetsd multi-view and multi band face recognitionIaetsd multi-view and multi band face recognition
Iaetsd multi-view and multi band face recognition
Iaetsd Iaetsd
 
Face Recognition System under Varying Lighting Conditions
Face Recognition System under Varying Lighting ConditionsFace Recognition System under Varying Lighting Conditions
Face Recognition System under Varying Lighting Conditions
IOSR Journals
 
Face and liveness detection with criminal identification using machine learni...
Face and liveness detection with criminal identification using machine learni...Face and liveness detection with criminal identification using machine learni...
Face and liveness detection with criminal identification using machine learni...
IAESIJAI
 
Real time multi face detection using deep learning
Real time multi face detection using deep learningReal time multi face detection using deep learning
Real time multi face detection using deep learning
Reallykul Kuul
 
Possibility fuzzy c means clustering for expression invariant face recognition
Possibility fuzzy c means clustering for expression invariant face recognitionPossibility fuzzy c means clustering for expression invariant face recognition
Possibility fuzzy c means clustering for expression invariant face recognition
IJCI JOURNAL
 
Deep hypersphere embedding for real-time face recognition
Deep hypersphere embedding for real-time face recognitionDeep hypersphere embedding for real-time face recognition
Deep hypersphere embedding for real-time face recognition
TELKOMNIKA JOURNAL
 
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
IRJET Journal
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
gerogepatton
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
gerogepatton
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
gerogepatton
 
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
ijtsrd
 
Face Recognition
Face RecognitionFace Recognition
Face Recognition
Saraj Sadanand
 
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUEFACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
IRJET Journal
 
Face Recognition & Detection Using Image Processing
Face Recognition & Detection Using Image ProcessingFace Recognition & Detection Using Image Processing
Face Recognition & Detection Using Image Processing
paperpublications3
 

Similar to How to Implement the Digital Medicine in the Future (20)

Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
Digital Future of the Surgery: Brining the Innovation of Digital Technology i...Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
Digital Future of the Surgery: Brining the Innovation of Digital Technology i...
 
Prediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep LearningPrediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep Learning
 
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
FACE MASK DETECTION AND COUNTER IN THINGSPEAK WITH EMAIL ALERT SYSTEM FOR COV...
 
Research Inventy : International Journal of Engineering and Science
Research Inventy : International Journal of Engineering and ScienceResearch Inventy : International Journal of Engineering and Science
Research Inventy : International Journal of Engineering and Science
 
IRJET- Breast Cancer Prediction using Deep Learning
IRJET-  	  Breast Cancer Prediction using Deep LearningIRJET-  	  Breast Cancer Prediction using Deep Learning
IRJET- Breast Cancer Prediction using Deep Learning
 
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
Study and Analysis of Novel Face Recognition Techniques using PCA, LDA and Ge...
 
Iaetsd multi-view and multi band face recognition
Iaetsd multi-view and multi band face recognitionIaetsd multi-view and multi band face recognition
Iaetsd multi-view and multi band face recognition
 
Face Recognition System under Varying Lighting Conditions
Face Recognition System under Varying Lighting ConditionsFace Recognition System under Varying Lighting Conditions
Face Recognition System under Varying Lighting Conditions
 
Face and liveness detection with criminal identification using machine learni...
Face and liveness detection with criminal identification using machine learni...Face and liveness detection with criminal identification using machine learni...
Face and liveness detection with criminal identification using machine learni...
 
Real time multi face detection using deep learning
Real time multi face detection using deep learningReal time multi face detection using deep learning
Real time multi face detection using deep learning
 
Possibility fuzzy c means clustering for expression invariant face recognition
Possibility fuzzy c means clustering for expression invariant face recognitionPossibility fuzzy c means clustering for expression invariant face recognition
Possibility fuzzy c means clustering for expression invariant face recognition
 
Deep hypersphere embedding for real-time face recognition
Deep hypersphere embedding for real-time face recognitionDeep hypersphere embedding for real-time face recognition
Deep hypersphere embedding for real-time face recognition
 
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
AI-based Mechanism to Authorise Beneficiaries at Covid Vaccination Camps usin...
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
 
A Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition SystemA Modified CNN-Based Face Recognition System
A Modified CNN-Based Face Recognition System
 
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
Face Recognition Based Attendance System with Auto Alert to Guardian using Ca...
 
Face Recognition
Face RecognitionFace Recognition
Face Recognition
 
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUEFACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
FACEMASK AND PHYSICAL DISTANCING DETECTION USING TRANSFER LEARNING TECHNIQUE
 
Face Recognition & Detection Using Image Processing
Face Recognition & Detection Using Image ProcessingFace Recognition & Detection Using Image Processing
Face Recognition & Detection Using Image Processing
 

More from Yoon Sup Choi

한국 원격의료 산업의 주요 이슈
한국 원격의료 산업의 주요 이슈한국 원격의료 산업의 주요 이슈
한국 원격의료 산업의 주요 이슈
Yoon Sup Choi
 
원격의료 시대의 디지털 치료제
원격의료 시대의 디지털 치료제원격의료 시대의 디지털 치료제
원격의료 시대의 디지털 치료제
Yoon Sup Choi
 
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
Yoon Sup Choi
 
디지털 헬스케어 파트너스 (DHP) 소개 자료
디지털 헬스케어 파트너스 (DHP) 소개 자료디지털 헬스케어 파트너스 (DHP) 소개 자료
디지털 헬스케어 파트너스 (DHP) 소개 자료
Yoon Sup Choi
 
[대한병리학회] 의료 인공지능 101: 병리를 중심으로
[대한병리학회] 의료 인공지능 101: 병리를 중심으로[대한병리학회] 의료 인공지능 101: 병리를 중심으로
[대한병리학회] 의료 인공지능 101: 병리를 중심으로
Yoon Sup Choi
 
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
Yoon Sup Choi
 
원격의료에 대한 생각, 그리고 그 생각에 대한 생각
원격의료에 대한 생각, 그리고 그 생각에 대한 생각원격의료에 대한 생각, 그리고 그 생각에 대한 생각
원격의료에 대한 생각, 그리고 그 생각에 대한 생각
Yoon Sup Choi
 
[C&C] 의료의 미래 디지털 헬스케어
[C&C] 의료의 미래 디지털 헬스케어[C&C] 의료의 미래 디지털 헬스케어
[C&C] 의료의 미래 디지털 헬스케어
Yoon Sup Choi
 
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
Yoon Sup Choi
 
디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약
Yoon Sup Choi
 
[365mc] 디지털 헬스케어: 의료의 미래
[365mc] 디지털 헬스케어: 의료의 미래[365mc] 디지털 헬스케어: 의료의 미래
[365mc] 디지털 헬스케어: 의료의 미래
Yoon Sup Choi
 
디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약
Yoon Sup Choi
 
[ASGO 2019] Artificial Intelligence in Medicine
[ASGO 2019] Artificial Intelligence in Medicine[ASGO 2019] Artificial Intelligence in Medicine
[ASGO 2019] Artificial Intelligence in Medicine
Yoon Sup Choi
 
글로벌 디지털 헬스케어 산업 및 규제 동향
글로벌 디지털 헬스케어 산업 및 규제 동향 글로벌 디지털 헬스케어 산업 및 규제 동향
글로벌 디지털 헬스케어 산업 및 규제 동향
Yoon Sup Choi
 
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
Yoon Sup Choi
 
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
Yoon Sup Choi
 
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
Yoon Sup Choi
 
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
Yoon Sup Choi
 
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
Yoon Sup Choi
 
디지털 의료가 '의료'가 될 때 (1/2)
디지털 의료가 '의료'가 될 때 (1/2)디지털 의료가 '의료'가 될 때 (1/2)
디지털 의료가 '의료'가 될 때 (1/2)
Yoon Sup Choi
 

More from Yoon Sup Choi (20)

한국 원격의료 산업의 주요 이슈
한국 원격의료 산업의 주요 이슈한국 원격의료 산업의 주요 이슈
한국 원격의료 산업의 주요 이슈
 
원격의료 시대의 디지털 치료제
원격의료 시대의 디지털 치료제원격의료 시대의 디지털 치료제
원격의료 시대의 디지털 치료제
 
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
[KNAPS] 포스트 코로나 시대, 제약 산업과 디지털 헬스케어
 
디지털 헬스케어 파트너스 (DHP) 소개 자료
디지털 헬스케어 파트너스 (DHP) 소개 자료디지털 헬스케어 파트너스 (DHP) 소개 자료
디지털 헬스케어 파트너스 (DHP) 소개 자료
 
[대한병리학회] 의료 인공지능 101: 병리를 중심으로
[대한병리학회] 의료 인공지능 101: 병리를 중심으로[대한병리학회] 의료 인공지능 101: 병리를 중심으로
[대한병리학회] 의료 인공지능 101: 병리를 중심으로
 
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
한국 디지털 헬스케어의 생존을 위한 규제 혁신에 대한 고언
 
원격의료에 대한 생각, 그리고 그 생각에 대한 생각
원격의료에 대한 생각, 그리고 그 생각에 대한 생각원격의료에 대한 생각, 그리고 그 생각에 대한 생각
원격의료에 대한 생각, 그리고 그 생각에 대한 생각
 
[C&C] 의료의 미래 디지털 헬스케어
[C&C] 의료의 미래 디지털 헬스케어[C&C] 의료의 미래 디지털 헬스케어
[C&C] 의료의 미래 디지털 헬스케어
 
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
포스트 코로나 시대, 혁신적인 디지털 헬스케어 기업의 조건
 
디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약
 
[365mc] 디지털 헬스케어: 의료의 미래
[365mc] 디지털 헬스케어: 의료의 미래[365mc] 디지털 헬스케어: 의료의 미래
[365mc] 디지털 헬스케어: 의료의 미래
 
디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약디지털 치료제, 또 하나의 신약
디지털 치료제, 또 하나의 신약
 
[ASGO 2019] Artificial Intelligence in Medicine
[ASGO 2019] Artificial Intelligence in Medicine[ASGO 2019] Artificial Intelligence in Medicine
[ASGO 2019] Artificial Intelligence in Medicine
 
글로벌 디지털 헬스케어 산업 및 규제 동향
글로벌 디지털 헬스케어 산업 및 규제 동향 글로벌 디지털 헬스케어 산업 및 규제 동향
글로벌 디지털 헬스케어 산업 및 규제 동향
 
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
인허가 이후에도 변화하는 AI/ML 기반 SaMD를 어떻게 규제할 것인가
 
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (상)
 
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
인공지능은 의료를 어떻게 혁신하는가 (2019년 7월) (하)
 
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
한국에서 혁신적인 디지털 헬스케어 스타트업이 탄생하려면
 
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
의료의 미래, 디지털 헬스케어 + 의료 시장의 특성
 
디지털 의료가 '의료'가 될 때 (1/2)
디지털 의료가 '의료'가 될 때 (1/2)디지털 의료가 '의료'가 될 때 (1/2)
디지털 의료가 '의료'가 될 때 (1/2)
 

Recently uploaded

anthelmintic-drugs.pptx pharmacology dep
anthelmintic-drugs.pptx pharmacology depanthelmintic-drugs.pptx pharmacology dep
anthelmintic-drugs.pptx pharmacology dep
sapnasirswal
 
JULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
JULY 2024 Oncology Cartoons by Dr Kanhu Charan PatroJULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
JULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
Kanhu Charan
 
Building a Strong Partnership with Your Medical Team
Building a Strong Partnership with Your Medical TeamBuilding a Strong Partnership with Your Medical Team
Building a Strong Partnership with Your Medical Team
bkling
 
SA Gastro Cure(pancreatic cancer treatment in india).pptx
SA Gastro Cure(pancreatic cancer treatment in india).pptxSA Gastro Cure(pancreatic cancer treatment in india).pptx
SA Gastro Cure(pancreatic cancer treatment in india).pptx
VinothKumar70905
 
BCBR MCQs with Answers.pdf for exam for NMC promotions
BCBR MCQs with Answers.pdf for exam for NMC promotionsBCBR MCQs with Answers.pdf for exam for NMC promotions
BCBR MCQs with Answers.pdf for exam for NMC promotions
sathya swaroop patnaik
 
Text Book of Nursing Concepts - Fundamental of Nursing
Text Book of Nursing Concepts - Fundamental of NursingText Book of Nursing Concepts - Fundamental of Nursing
Text Book of Nursing Concepts - Fundamental of Nursing
BP KOIRALA INSTITUTE OF HELATH SCIENCS,, NEPAL
 
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
NephroTube - Dr.Gawad
 
Basic life support and its management for students
Basic life support and its management  for studentsBasic life support and its management  for students
Basic life support and its management for students
MayarHamed3
 
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
PVI, PeerView Institute for Medical Education
 
medical law and ethics presentation .ppt
medical law and ethics presentation .pptmedical law and ethics presentation .ppt
medical law and ethics presentation .ppt
PseudoPocket
 
Top 10 Habits for Longevity [Biohacker Summit 2024]
Top 10 Habits for Longevity [Biohacker Summit 2024]Top 10 Habits for Longevity [Biohacker Summit 2024]
Top 10 Habits for Longevity [Biohacker Summit 2024]
Olli Sovijärvi
 
selllllllllllllllllllllllllllllllllllllllllllllll.pptx
selllllllllllllllllllllllllllllllllllllllllllllll.pptxselllllllllllllllllllllllllllllllllllllllllllllll.pptx
selllllllllllllllllllllllllllllllllllllllllllllll.pptx
Joebest8
 
PCF-Assessment-Tool_Policy-Guide (1).pdf
PCF-Assessment-Tool_Policy-Guide (1).pdfPCF-Assessment-Tool_Policy-Guide (1).pdf
PCF-Assessment-Tool_Policy-Guide (1).pdf
AbHermoso
 
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptxHow to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
AmandaChou9
 
Clinical examination of- CRANIAL.- nerves
Clinical examination of- CRANIAL.- nervesClinical examination of- CRANIAL.- nerves
Clinical examination of- CRANIAL.- nerves
DrpoonamHealthclinic
 
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
Anindya Das Adhikary
 
Respiratory system at glance- Neonatology
Respiratory system at glance- NeonatologyRespiratory system at glance- Neonatology
Respiratory system at glance- Neonatology
Dr. Habibur Rahim
 
Article - Design and evaluation of novel inhibitors for the treatment of clea...
Article - Design and evaluation of novel inhibitors for the treatment of clea...Article - Design and evaluation of novel inhibitors for the treatment of clea...
Article - Design and evaluation of novel inhibitors for the treatment of clea...
Trustlife
 
Types of Hypoxia, Hypercapnia, and Cyanosis
Types of Hypoxia, Hypercapnia, and CyanosisTypes of Hypoxia, Hypercapnia, and Cyanosis
Types of Hypoxia, Hypercapnia, and Cyanosis
MedicoseAcademics
 
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
Niranjan Chavan
 

Recently uploaded (20)

anthelmintic-drugs.pptx pharmacology dep
anthelmintic-drugs.pptx pharmacology depanthelmintic-drugs.pptx pharmacology dep
anthelmintic-drugs.pptx pharmacology dep
 
JULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
JULY 2024 Oncology Cartoons by Dr Kanhu Charan PatroJULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
JULY 2024 Oncology Cartoons by Dr Kanhu Charan Patro
 
Building a Strong Partnership with Your Medical Team
Building a Strong Partnership with Your Medical TeamBuilding a Strong Partnership with Your Medical Team
Building a Strong Partnership with Your Medical Team
 
SA Gastro Cure(pancreatic cancer treatment in india).pptx
SA Gastro Cure(pancreatic cancer treatment in india).pptxSA Gastro Cure(pancreatic cancer treatment in india).pptx
SA Gastro Cure(pancreatic cancer treatment in india).pptx
 
BCBR MCQs with Answers.pdf for exam for NMC promotions
BCBR MCQs with Answers.pdf for exam for NMC promotionsBCBR MCQs with Answers.pdf for exam for NMC promotions
BCBR MCQs with Answers.pdf for exam for NMC promotions
 
Text Book of Nursing Concepts - Fundamental of Nursing
Text Book of Nursing Concepts - Fundamental of NursingText Book of Nursing Concepts - Fundamental of Nursing
Text Book of Nursing Concepts - Fundamental of Nursing
 
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
Hemodialysis: Chapter 9, Arteriovenous Fistula and Graft: Basics, Creation, U...
 
Basic life support and its management for students
Basic life support and its management  for studentsBasic life support and its management  for students
Basic life support and its management for students
 
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
Stepping Forward to Transform MCL Management: Guidance on the Selection and U...
 
medical law and ethics presentation .ppt
medical law and ethics presentation .pptmedical law and ethics presentation .ppt
medical law and ethics presentation .ppt
 
Top 10 Habits for Longevity [Biohacker Summit 2024]
Top 10 Habits for Longevity [Biohacker Summit 2024]Top 10 Habits for Longevity [Biohacker Summit 2024]
Top 10 Habits for Longevity [Biohacker Summit 2024]
 
selllllllllllllllllllllllllllllllllllllllllllllll.pptx
selllllllllllllllllllllllllllllllllllllllllllllll.pptxselllllllllllllllllllllllllllllllllllllllllllllll.pptx
selllllllllllllllllllllllllllllllllllllllllllllll.pptx
 
PCF-Assessment-Tool_Policy-Guide (1).pdf
PCF-Assessment-Tool_Policy-Guide (1).pdfPCF-Assessment-Tool_Policy-Guide (1).pdf
PCF-Assessment-Tool_Policy-Guide (1).pdf
 
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptxHow to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
How to Relieve Prostate Congestion- Here are some Effective Strategies.pptx
 
Clinical examination of- CRANIAL.- nerves
Clinical examination of- CRANIAL.- nervesClinical examination of- CRANIAL.- nerves
Clinical examination of- CRANIAL.- nerves
 
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
Amygdala Medi-Trivia Quiz (Prelims) | FAQ 2024
 
Respiratory system at glance- Neonatology
Respiratory system at glance- NeonatologyRespiratory system at glance- Neonatology
Respiratory system at glance- Neonatology
 
Article - Design and evaluation of novel inhibitors for the treatment of clea...
Article - Design and evaluation of novel inhibitors for the treatment of clea...Article - Design and evaluation of novel inhibitors for the treatment of clea...
Article - Design and evaluation of novel inhibitors for the treatment of clea...
 
Types of Hypoxia, Hypercapnia, and Cyanosis
Types of Hypoxia, Hypercapnia, and CyanosisTypes of Hypoxia, Hypercapnia, and Cyanosis
Types of Hypoxia, Hypercapnia, and Cyanosis
 
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
STRATEGIES FOR RATIONALISING/REDUCING CAESAREAN SECTION RATE BY USE OF "SION ...
 

How to Implement the Digital Medicine in the Future

  • 1. Sungkyunkwan University Department of Human ICT Convergence Yoon Sup Choi, Ph.D. How to implement the digital medicine in the future : measure, collect and interpret patient-generated data
  • 2. The Convergence of IT, BT and Medicine
  • 7. What is most important factor in digital medicine?
  • 8. “Data! Data! Data!” he cried.“I can’t make bricks without clay!” - Sherlock Holmes,“The Adventure of the Copper Beeches”
  • 9. Three Steps to Implement Digital Medicine • Step 1. Measure the Data • Step 2. Collect the Data • Step 3. Insight from the Data
  • 10. Step 1. Measure the Data
  • 11. Smartphone: the origin of healthcare innovation
  • 12. 2013? The election of Pope Benedict The Election of Pope Francis
  • 13. The Election of Pope Francis The Election of Pope Benedict
  • 18. PEEK (portable eye examination kit) http://www.peekvision.org
  • 19. OScan: oral cancer detection
  • 34. C8 Medisensor: non-invasive blood glucose sensor
  • 35. Withings Wireless Blood Pressure Monitor
  • 36. Huinno: Cuff-less Blood Pressure Monitor
  • 42. Results within 6-8 weeksA little spit is all it takes! DTC Genetic TestingDirect-To-Consumer
  • 43. 120 Disease Risk 21 Drug Response 49 Carrier Status 57Traits $99
  • 48. Traits 음주 후 얼굴이 붉어지는가 쓴 맛을 감지할 수 있나 귀지 유형 눈 색깔 곱슬머리 여부 유당 분해 능력 말라리아 저항성 대머리가 될 가능성 근육 퍼포먼스 혈액형 노로바이러스 저항성 HIV 저항성 흡연 중독 가능성
  • 53. Step1. Measure the Data • With your smartphone • With wearable devices (connected to smartphone) • Personal genome analysis ... without even going to the hospital!
  • 54. Step 2. Collect the Data
  • 61. Epic MyChart App Epic EHRDatabaseDexcom App Withings App Dexcom CGM Nike+ Patients Device/Apps HealthKit EHR Hospital Whitings + • Data stored in DB on the iPhone (, not mirroring to the cloud) • Consumer controls what data goes in/out, privacy level • HealthKit connects/direct devices, store data based on privacy rules Apple Watch iPhone
  • 65. Without cloud computing, we cannot collect the real-time big data from the patients
  • 67. Practice Fusion, an EMR based on the cloud
  • 68. Step 3. Insight from the Data
  • 70. How to Analyze and Interpret the Big Data?
  • 71. and/or Two ways to get insights from the big data
  • 72. Hospitals in the future: Data Analysis Center
  • 73. Doctors in the future: Data Scientists
  • 74. No choice but to bring AI into the medicine
  • 77. “AliveCor has received an additional FDA 510(k) clearance, this time for an algorithm that allows its smartphone ECG to detect atrial fibrillation with high accuracy.” “the algorithm has a 100 percent sensitivity (it never returns a false negative) and a 97 percent specificity (it returns false positives about 3 percent of the time). For obvious reasons, the algorithm was designed to err on the side of false positives”
  • 78. DeepFace: Closing the Gap to Human-Level Performance in FaceVerification Taigman,Y. et al. (2014). DeepFace: Closing the Gap to Human-Level Performance in FaceVerification, CVPR’14. Figure 2. Outline of the DeepFace architecture. A front-end of a single convolution-pooling-convolution filtering on the rectified input, followed by three locally-connected layers and two fully-connected layers. Colors illustrate feature maps produced at each layer. The net includes more than 120 million parameters, where more than 95% come from the local and fully connected layers. very few parameters. These layers merely expand the input into a set of simple local features. The subsequent layers (L4, L5 and L6) are instead lo- cally connected [13, 16], like a convolutional layer they ap- ply a filter bank, but every location in the feature map learns a different set of filters. Since different regions of an aligned image have different local statistics, the spatial stationarity The goal of training is to maximize the probability of the correct class (face id). We achieve this by minimiz- ing the cross-entropy loss for each training sample. If k is the index of the true label for a given input, the loss is: L = log pk. The loss is minimized over the parameters by computing the gradient of L w.r.t. the parameters and by updating the parameters using stochastic gradient de- Human: 95% vs. DeepFace in Facebook: 97.35% Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)
  • 79. FaceNet:A Unified Embedding for Face Recognition and Clustering Schroff, F. et al. (2015). FaceNet:A Unified Embedding for Face Recognition and Clustering Human: 95% vs. FaceNet of Google: 99.63% Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people) False accept False reject s. This shows all pairs of images that were on LFW. Only eight of the 13 errors shown the other four are mislabeled in LFW. on Youtube Faces DB ge similarity of all pairs of the first one our face detector detects in each video. False accept False reject Figure 6. LFW errors. This shows all pairs of images that were incorrectly classified on LFW. Only eight of the 13 errors shown here are actual errors the other four are mislabeled in LFW. 5.7. Performance on Youtube Faces DB We use the average similarity of all pairs of the first one hundred frames that our face detector detects in each video. This gives us a classification accuracy of 95.12%±0.39. Using the first one thousand frames results in 95.18%. Compared to [17] 91.4% who also evaluate one hundred frames per video we reduce the error rate by almost half. DeepId2+ [15] achieved 93.2% and our method reduces this error by 30%, comparable to our improvement on LFW. 5.8. Face Clustering Our compact embedding lends itself to be used in order to cluster a users personal photos into groups of people with the same identity. The constraints in assignment imposed by clustering faces, compared to the pure verification task, lead to truly amazing results. Figure 7 shows one cluster in a users personal photo collection, generated using agglom- erative clustering. It is a clear showcase of the incredible invariance to occlusion, lighting, pose and even age. Figure 7. Face Clustering. Shown is an exemplar cluster for one user. All these images in the users personal photo collection were clustered together. 6. Summary We provide a method to directly learn an embedding into an Euclidean space for face verification. This sets it apart from other methods [15, 17] who use the CNN bottleneck layer, or require additional post-processing such as concate- nation of multiple models and PCA, as well as SVM clas- sification. Our end-to-end training both simplifies the setup and shows that directly optimizing a loss relevant to the task at hand improves performance. Another strength of our model is that it only requires False accept False reject Figure 6. LFW errors. This shows all pairs of images that were incorrectly classified on LFW. Only eight of the 13 errors shown here are actual errors the other four are mislabeled in LFW. 5.7. Performance on Youtube Faces DB We use the average similarity of all pairs of the first one hundred frames that our face detector detects in each video. This gives us a classification accuracy of 95.12%±0.39. Using the first one thousand frames results in 95.18%. Compared to [17] 91.4% who also evaluate one hundred frames per video we reduce the error rate by almost half. DeepId2+ [15] achieved 93.2% and our method reduces this error by 30%, comparable to our improvement on LFW. 5.8. Face Clustering Our compact embedding lends itself to be used in order to cluster a users personal photos into groups of people with the same identity. The constraints in assignment imposed by clustering faces, compared to the pure verification task, Figure 7. Face Clustering. Shown is an exemplar cluster for one user. All these images in the users personal photo collection were clustered together. 6. Summary We provide a method to directly learn an embedding into an Euclidean space for face verification. This sets it apart from other methods [15, 17] who use the CNN bottleneck layer, or require additional post-processing such as concate- nation of multiple models and PCA, as well as SVM clas-
  • 80. Constructing higher-level contextual/relational features: Relationships between epithelial nuclear neighbors Relationships between morphologically regular and irregular nuclei Relationships between epithelial and stromal objects Relationships between epithelial nuclei and cytoplasm Characteristics of stromal nuclei and stromal matrix Characteristics of epithelial nuclei and epithelial cytoplasm Building an epithelial/stromal classifier: Epithelial vs.stroma classifier Epithelial vs.stroma classifier B Basic image processing and feature construction: H&E image Image broken into superpixels Nuclei identified within each superpixel A Relationships of contiguous epithelial regions with underlying nuclear objects Learning an image-based model to predict survival Processed images from patients Processed images from patients C D onNovember17,2011stm.sciencemag.orgwnloadedfrom TMAs contain 0.6-mm-diameter cores (median of two cores per case) that represent only a small sample of the full tumor. We acquired data from two separate and independent cohorts: Nether- lands Cancer Institute (NKI; 248 patients) and Vancouver General Hospital (VGH; 328 patients). Unlike previous work in cancer morphom- etry (18–21), our image analysis pipeline was not limited to a predefined set of morphometric features selected by pathologists. Rather, C-Path measures an extensive, quantitative feature set from the breast cancer epithelium and the stro- ma (Fig. 1). Our image processing system first performed an automated, hierarchical scene seg- mentation that generated thousands of measure- ments, including both standard morphometric descriptors of image objects and higher-level contextual, relational, and global image features. The pipeline consisted of three stages (Fig. 1, A to C, and tables S8 and S9). First, we used a set of processing steps to separate the tissue from the background, partition the image into small regions of coherent appearance known as superpixels, find nuclei within the superpixels, and construct Constructing higher-level contextual/relational features: Relationships between epithelial nuclear neighbors Relationships between morphologically regular and irregular nuclei Relationships between epithelial and stromal objects Relationships between epithelial nuclei and cytoplasm Characteristics of stromal nuclei and stromal matrix Characteristics of epithelial nuclei and epithelial cytoplasm Epithelial vs.stroma classifier Epithelial vs.stroma classifier Relationships of contiguous epithelial regions with underlying nuclear objects Learning an image-based model to predict survival Processed images from patients alive at 5 years Processed images from patients deceased at 5 years L1-regularized logisticregression modelbuilding 5YS predictive model Unlabeled images Time P(survival) C D Identification of novel prognostically important morphologic features basic cellular morphologic properties (epithelial reg- ular nuclei = red; epithelial atypical nuclei = pale blue; epithelial cytoplasm = purple; stromal matrix = green; stromal round nuclei = dark green; stromal spindled nuclei = teal blue; unclassified regions = dark gray; spindled nuclei in unclassified regions = yellow; round nuclei in unclassified regions = gray; background = white). (Left panel) After the classification of each image object, a rich feature set is constructed. (D) Learning an image-based model to predict survival. Processed images from patients alive at 5 years after surgery and from patients deceased at 5 years after surgery were used to construct an image-based prog- nostic model. After construction of the model, it was applied to a test set of breast cancer images (not used in model building) to classify patients as high or low risk of death by 5 years. www.ScienceTranslationalMedicine.org 9 November 2011 Vol 3 Issue 108 108ra113 2 onNovember17,2011stm.sciencemag.orgDownloadedfrom Digital Pathologist Sci Transl Med. 2011 Nov 9;3(108):108ra113
  • 81. Digital Pathologist Sci Transl Med. 2011 Nov 9;3(108):108ra113 Top stromal features associated with survival. primarily characterizing epithelial nuclear characteristics, such as size, color, and texture (21, 36). In contrast, after initial filtering of im- ages to ensure high-quality TMA images and training of the C-Path models using expert-derived image annotations (epithelium and stroma labels to build the epithelial-stromal classifier and survival time and survival status to build the prognostic model), our image analysis system is automated with no manual steps, which greatly in- creases its scalability. Additionally, in contrast to previous approaches, our system measures thousands of morphologic descriptors of diverse identification of prognostic features whose significance was not pre- viously recognized. Using our system, we built an image-based prognostic model on the NKI data set and showed that in this patient cohort the model was a strong predictor of survival and provided significant additional prognostic information to clinical, molecular, and pathological prog- nostic factors in a multivariate model. We also demonstrated that the image-based prognostic model, built using the NKI data set, is a strong prognostic factor on another, independent data set with very different SD of the ratio of the pixel intensity SD to the mean intensity for pixels within a ring of the center of epithelial nuclei A The sum of the number of unclassified objects SD of the maximum blue pixel value for atypical epithelial nuclei Maximum distance between atypical epithelial nuclei B C D Maximum value of the minimum green pixel intensity value in epithelial contiguous regions Minimum elliptic fit of epithelial contiguous regions SD of distance between epithelial cytoplasmic and nuclear objects Average border between epithelial cytoplasmic objects E F G H Fig. 5. Top epithelial features. The eight panels in the figure (A to H) each shows one of the top-ranking epithelial features from the bootstrap anal- ysis. Left panels, improved prognosis; right panels, worse prognosis. (A) SD of the (SD of intensity/mean intensity) for pixels within a ring of the center of epithelial nuclei. Left, relatively consistent nuclear intensity pattern (low score); right, great nuclear intensity diversity (high score). (B) Sum of the number of unclassified objects. Red, epithelial regions; green, stromal re- gions; no overlaid color, unclassified region. Left, few unclassified objects (low score); right, higher number of unclassified objects (high score). (C) SD of the maximum blue pixel value for atypical epithelial nuclei. Left, high score; right, low score. (D) Maximum distance between atypical epithe- lial nuclei. Left, high score; right, low score. (Insets) Red, atypical epithelial nuclei; black, typical epithelial nuclei. (E) Minimum elliptic fit of epithelial contiguous regions. Left, high score; right, low score. (F) SD of distance between epithelial cytoplasmic and nuclear objects. Left, high score; right, low score. (G) Average border between epithelial cytoplasmic objects. Left, high score; right, low score. (H) Maximum value of the minimum green pixel intensity value in epithelial contiguous regions. Left, low score indi- cating black pixels within epithelial region; right, higher score indicating presence of epithelial regions lacking black pixels. onNovember17,2011stm.sciencemag.orgDownloadedfrom and stromal matrix throughout the image, with thin cords of epithe- lial cells infiltrating through stroma across the image, so that each stromal matrix region borders a relatively constant proportion of ep- ithelial and stromal regions. The stromal feature with the second largest coefficient (Fig. 4B) was the sum of the minimum green in- tensity value of stromal-contiguous regions. This feature received a value of zero when stromal regions contained dark pixels (such as inflammatory nuclei). The feature received a positive value when stromal objects were devoid of dark pixels. This feature provided in- formation about the relationship between stromal cellular composi- tion and prognosis and suggested that the presence of inflammatory cells in the stroma is associated with poor prognosis, a finding con- sistent with previous observations (32). The third most significant stromal feature (Fig. 4C) was a measure of the relative border between spindled stromal nuclei to round stromal nuclei, with an increased rel- ative border of spindled stromal nuclei to round stromal nuclei asso- ciated with worse overall survival. Although the biological underpinning of this morphologic feature is currently not known, this analysis sug- gested that spatial relationships between different populations of stro- mal cell types are associated with breast cancer progression. Reproducibility of C-Path 5YS model predictions on samples with multiple TMA cores For the C-Path 5YS model (which was trained on the full NKI data set), we assessed the intrapatient agreement of model predictions when predictions were made separately on each image contributed by pa- tients in the VGH data set. For the 190 VGH patients who contributed two images with complete image data, the binary predictions (high or low risk) on the individual images agreed with each other for 69% (131 of 190) of the cases and agreed with the prediction on the aver- aged data for 84% (319 of 380) of the images. Using the continuous prediction score (which ranged from 0 to 100), the median of the ab- solute difference in prediction score among the patients with replicate images was 5%, and the Spearman correlation among replicates was 0.27 (P = 0.0002) (fig. S3). This degree of intrapatient agreement is only moderate, and these findings suggest significant intrapatient tumor heterogeneity, which is a cardinal feature of breast carcinomas (33–35). Qualitative visual inspection of images receiving discordant scores suggested that intrapatient variability in both the epithelial and the stromal components is likely to contribute to discordant scores for the individual images. These differences appeared to relate both to the proportions of the epithelium and stroma and to the appearance of the epithelium and stroma. Last, we sought to analyze whether sur- vival predictions were more accurate on the VGH cases that contributed multiple cores compared to the cases that contributed only a single core. This analysis showed that the C-Path 5YS model showed signif- icantly improved prognostic prediction accuracy on the VGH cases for which we had multiple images compared to the cases that con- tributed only a single image (Fig. 7). Together, these findings show a significant degree of intrapatient variability and indicate that increased tumor sampling is associated with improved model performance. DISCUSSION Heat map of stromal matrix objects mean abs.diff to neighbors H&E image separated into epithelial and stromal objects A B C Worse prognosis Improved prognosis Improved prognosis Improved prognosis Worse prognosis Worse prognosis Fig. 4. Top stromal features associated with survival. (A) Variability in ab- solute difference in intensity between stromal matrix regions and neigh- bors. Top panel, high score (24.1); bottom panel, low score (10.5). (Insets) Top panel, high score; bottom panel; low score. Right panels, stromal matrix objects colored blue (low), green (medium), or white (high) according to each object’s absolute difference in intensity to neighbors. (B) Presence R E S E A R C H A R T I C L E onNovember17,2011stm.sciencemag.orgDownloadedfrom Top epithelial features.The eight panels in the figure (A to H) each shows one of the top-ranking epithelial features from the bootstrap anal- ysis. Left panels, improved prognosis; right panels, worse prognosis.
  • 91. + Integration of Health Data and Genomic Data
  • 92. + +
  • 95. • Apple HealthKit • Fitbit Data • Personal Genome Data • GPS + Personalized Healthcare Advices
  • 98. Three Steps to Implement Digital Medicine • Step 1. Measure the Data • Step 2. Collect the Data • Step 3. Insight from the Data
  • 100. Feedback/Questions • Email: yoonsup.choi@gmail.com • Blog: http://www.yoonsupchoi.com • Facebook: Yoon Sup Choi