The mass of collected data opens the door to much more accurate analyzes of the healthcare system – with the purpose of optimizing it – and to a shift from preventive to predictive medicine. With big data providing and increase tools to prevent and predict pathologies, the data scientist role emerges in the clinical space, suggesting an evolution of the medical professionals.
BPAC WITH UFSBI GENERAL PRESENTATION 18_05_2017-1.pptx
2017 China IoT Conference, Dec.6-7, 2017 | Shenzhen
1. The future of wearable devices.
How AI can revolutionize approach
to medical device development
Ron Fridman
AI lecture CEO & FounderPartner
2. Even Elon Musk talks about Ukraine
Source: http://scienews.com/space/1271-elon-musk-said-that-the-
ukrainian-rocket-zenit-the-best-after-his-falc.html
8. What is the keyword in the future of
medical devices?
Soures: https://www.economist.com/news/leaders/21721656-data-economy-demands-new-approach-antitrust-rules-worlds-most-
valuable-resource
13. Our R&D Projects
Arrhythmia detection Parkinson disease
detection
Emotion detection
using ECG
Biometric identification
using ECG
Activity tracking
and fitness optimization
AI - based EEG
analysis
Detection of non
heart-related diseases
14. FDA is ready to certify even AI
https://www.forbes.com/sites/bernardmarr/2017/01/20/
first-fda-approval-for-clinical-cloud-based-deep-learning-in-
healthcare/#509f4f1d161c
15. What do these people have in common?
Vice President
Joe Biden
Basketball
Hall of Famer
Larry Bird
NASCAR driver
Michael Walltrip
Musician
Barry Manilow
My uncle
Alex
Source: http://www.daily-journal.com/opinion/columnists/local/discovering-a-fib-common-awkward-
potentially-deadly/article_62e5ddcd-a56b-587b-9269-1421fa2d7122.html
They are among the estimated 3.2 Americans who have a
heart condition called atrial fibrillation or “AF”.
16. 9 of 10 heart strokes could be predicted
http://merisight.com/
3.2M
U.S people diagnosed with
AF + 1M undiagnosed
460.000
AF inpatient
Hospitalization in 2012
+5-10%
Yearly growth in AF population
1 in 4 people over forty
develops AF
17. Our approach to
arrhythmia detection
• A prototype of AI for automated arrhythmia
detection and tested it on different datasets.
• An accuracy of 95% on detecting atrial
fibrillation.
• 20 seconds long ECG records as an input to
our deep neural network.
18. Parkinson’s disease
statistics
The exact cause of
Parkinson’s disease
is unknown.
There is no test to
diagnose Parkinson’s
disease.
Source: https://www.michaeljfox.org/page.html?what-is-
parkinsons-infographic
19. Main functions of the system
Detection of Parkinson’s
disease on early stage
Analysis of the patient's
condition
Interaction between
patient and doctor
Automatic lifestyle
recommendations that can
improve patient’s condition
20. Emotion Detection
According to the most popular view on
humans emotions, they are all
characterized by 2 parameters:
• Arousal - may be explained as “how
powerful is an emotion”
• Valence - is emotion positive or negative
https://www.researchgate.net/profile/Yi-hsuan_Yang/publication/
254004106/figure/fig1/AS:298208942149638@1448109960909/
Fig-2-The-2D-valence-arousal-emotion-space-Russell-1980-the-position-of-the.png
21. Wearables can predict even fertility days
Standard fertility awareness methods:
Standard days Body temperature Urine or another test
Not accurate Not comfortable Not comfortable at all
22. Fertility tracking features:
- Electrocardiogram (ECG)
- Sleep quality
- Stress level
- Vital signs
our AI based solution:
- shows better accuracy | >95%
- more comfortable | wristband
- bigger picture of the health | via ECG
Our novel approach to prediction fertility days