SlideShare a Scribd company logo
Imaging and Vision
Pathfinding
Perry Lea
ACM Distinguished Lectures
Floating Point
2D
Graphics
3D
Graphics
Vision Computational
Photography
Physics Kernel Floating Point
Requirement
X Color Space
Conversion
Fixed Point
X Gaussian Blur Fixed Point
X Sobel Edge
Detection
Fixed Point
X Bilateral Filters Fixed Point
X Bilinear
Interpolation
Fixed Point
X Bicubic
Interpolation
Half or Single Precision
X Image Signal
Processor
Fixed Point
X X Exposure
Compensation
Single Precision
X X Image Blending Fixed Point
X X Scaling Fixed (for binary scaling)
X Texture Mapping Fixed Point
X Pixel Shading Single / Double Precision
X Z-Buffer Depth Test Single
X Compositing Fixed Point or Half
Precision
X Ray Tracing Single Precision
X 3D Vertex Shading Double Precision
X Fluid Dynamics Single / Double Precision
X JPEG Compression Fixed Point2
Vision
Vision Segments
4
ADAS and Automotive
Medical Imaging
Consumer Electronics and GamingIndustrial Automation &
Robotics
Security, Surveillance,
Intelligence, Defense
Facial Recognition
Why Vision
5
Vision Market
6
Tractica Research: 42% CAGR, $33B market by 2019
Market to Market: 22.6% CAGR, $22.2B market by 2020
Human Vision
| February 11,
2018
| Micron Confidential
7
What do you see here?
Do you see lines between the
circles?
 Guess what: there are none.
Rule 1: Sensory input does not
contain enough information to
explain our perception
What did you just see?
 Did you see the people on the bridge?
 Did you see the church?
 Did you see the tunnel?
Rule 2: There is too much sensory
input to include in our coherent
perceptions at any single moment
8
Human Visual Dataflow
Human vision interprets
images bottom up and top
down:
Bottom Up: Based on raw
sensory data (pixels)
Top down: based on feature
extraction
Find the Target
9
Human Brain Visual System
from Ganglion to Cortex
How Human Vision Works
Humans are born with a nearly fully
developed vision system
Cortical pathways are reinforced and
restructured within the 1st year of
development.
Vision starts at ganglion
cells and follows
the optical nerve.
Some receptors will excite with light
intensity, some will inhibit activity.
1
0
Feature Extraction
When a collection of
photoreceptors are organized
into a center-surround field,
the brain can easily perceive
light and dark regions.
Edges force ganglion cells to
deliver reinforced or
diminished signals.
Visual System does an
extraordinary job at throwing
away information.
1
1
Ganglion Cell Signal Strength
Computer Vision
Vision Principles
SIFT in 6 slides!
Just as the human brain perceives image data top-down and bottom-up,
so are typical vision algorithms.
 Features are “interesting” parts of an image and we will rely on the same
edges, corners, and ridges. To be useful, feature points must:
Be numerous
Be repeatable
Represent orientation
and scale
Be fast to extract
and match
1
3
Typical Feature Extraction Algorithm
Detector
 Find Scale Space
Extrema
 Keypoint Localization
Improve keypoints and
throw out bad ones.
Descriptor
 Orientation Assignment
(remove effects of
rotation and scale)
 Create Descriptor
Use histograms of
orientations
1
4
Lens
Lens
Correction
White
Balance
Noise
Reduction
Demosaic
Color
Correction
Tone
Mapping
Sharpening
Gamma
Correction
3A Stats
RGB2YUV Scaler
DRAM
Image Signal Processor (Front End)
Feature Extraction (Back End)
12 MegaPixel Image (RAW10=15 MB to
37MB. @30 fps = 450 MB/s)
Preprocess Scan Image
Filter Feature
Locations
Generate
Signature
Post Process
Descriptors
Finding Scale Space
Finds keypoints in image.
Image is convolved at different
scales (variant of blob detection)
Best way to do this is a Laplacian of Gaussian:
 But a LoG is really computationally expensive (hmmm)
 So we’ll cheat and do a Difference of Gaussian Blurs:
 Convolved images are grouped by “octaves” which is simply the scale at that
point. We convolve a certain number of images per octave k
 Take the difference of the convolved images k per octave.
1
5
Finding Scale Space
Find Extreme
Choose all extrema within a 3x3x3
neighborhood
 This is done by comparing each pixel in the
DoG images to its eight neighbors at the
same scale and nine corresponding
neighboring pixels in each of the neighboring
scales. If the pixel value is the maximum or
minimum among all compared pixels, it is
selected as a candidate keypoint.
1
6
Keypoint Localization
Scale space extrema produce too many
candidates.
Minimize:
 Use Taylor series expansion to get
true extrema
Reject:
 Points with bad contrast
 Points with strong edge response in 1 direction
1
7
Orientation Assignment
Remove effects of rotation
Create a gradient of histograms (36 bins)
Weighted by magnitude of Gaussian Window
Any peak within 80% of highest is a new keypoint
Parabola a parabola is fit to the 3 histograms closest to each peak
1
8
Keypoint Descriptor
We now want to compute a descriptor for each keypoint to make
them distinctive with various illuminations, 3D views, etc.
Similar to human biological vision
Neurons respond to gradients at certain frequencies
4x4 gradient window with a histogram of 4x4 samples per
window = 4x4x8 = 128 feature vectors
1
9
Lighting gains will
not affect descriptors
Feature Detection Algorithms
Edge Detection:
Canny, Sobel, Prewitt,
Differential
Corner Detectors:
Harris, FAST, SUSAN
Blob Detectors:
Laplacian of Gaussian
Difference of Gaussian
Determinant of Gaussian
2
0
 Transforms:
– Ridge, Hough, Structural
Tensor
 Affine Invariants
– Affine shape adapter
– Harris Affine
– Hessian Affine
 Feature Descriptors
– SIFT, SURF, GLOH, HOG,
BRIEF, ORB, BRISK, FREAK
Other Vision Challenges
Segmentation
Meaningful partitioning of
image/video into non-overlapping
regions and subvolumes. Ability to
handle multi-modal data of varying
complexity
2
1
Color Image Segmentation Output
Original Image courtesy of
University of California at Berkeley
Courtesy RIT
Other Vision Challenges
Super Resolution
Utilizing multiple images of a given scene to obtain a high
resolution image with improved image quality
2
2
Other Vision Challenges
Hierarchical Scale Space
Using information at various scales to determine the semantic
structure of an image. Utilize probabilistic modeling of an image
content to build a dynamic hierarchical tree for high resolution
remote sensing.
2
3
Courtesy RIT
Other Vision Challenges
Computational Photography
2
4
Computational photography combines
plentiful computing, digital sensors,
modern optics, actuators, and smart lights
to escape the limitations of traditional film
cameras and enables novel imaging
applications. Unbounded dynamic range,
variable focus, resolution, and depth of
field, hints about shape, reflectance, and
lighting, and new interactive forms of
photos that are partly snapshots and partly
videos are just some of the new
applications found in Computational
Photography.
• Light Field Arrays
• Massive Image Stitching/Warping
• Computational Optics
• Holographic Imaging

More Related Content

What's hot

Computer vision
Computer visionComputer vision
Computer vision
ghufranAlkaaby
 
Open CV - 電腦怎麼看世界
Open CV - 電腦怎麼看世界Open CV - 電腦怎麼看世界
Open CV - 電腦怎麼看世界
Tech Podcast Night
 
Face detection ppt
Face detection pptFace detection ppt
Face detection ppt
Pooja R
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
ArtiKhanchandani
 
Digital Image Processing and Edge Detection
Digital Image Processing and Edge DetectionDigital Image Processing and Edge Detection
Digital Image Processing and Edge Detection
Seda Yalçın
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
Nitin Sharma
 
Computer vision suprim regmi
Computer vision suprim regmi Computer vision suprim regmi
Computer vision suprim regmi
Suprim Regmi
 
Image Processing and Computer Vision
Image Processing and Computer VisionImage Processing and Computer Vision
Image Processing and Computer Vision
Silicon Mentor
 
AI Computer vision
AI Computer visionAI Computer vision
AI Computer vision
Kashafnaz2
 
Application of edge detection
Application of edge detectionApplication of edge detection
Application of edge detection
Naresh Biloniya
 
Computer vision for interactive computer graphics
Computer vision for interactive computer graphicsComputer vision for interactive computer graphics
Computer vision for interactive computer graphics
Shah Alam Sabuj
 
Ai lecture 03 computer vision
Ai lecture 03 computer visionAi lecture 03 computer vision
Ai lecture 03 computer vision
Ahmad sohail Kakar
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
Ameer Mohamed Rajah
 
Computer Vision - Artificial Intelligence
Computer Vision - Artificial IntelligenceComputer Vision - Artificial Intelligence
Computer Vision - Artificial Intelligence
ACM-KU
 
computer vision & Opencv intro
computer vision & Opencv intro computer vision & Opencv intro
computer vision & Opencv intro
チュニジア の自由
 
A350111
A350111A350111
Computer vision and robotics
Computer vision and roboticsComputer vision and robotics
Computer vision and robotics
Biniam Asnake
 
Interactive full body motion capture using infrared sensor network
Interactive full body motion capture using infrared sensor networkInteractive full body motion capture using infrared sensor network
Interactive full body motion capture using infrared sensor network
ijcga
 
Interactive Full-Body Motion Capture Using Infrared Sensor Network
Interactive Full-Body Motion Capture Using Infrared Sensor Network  Interactive Full-Body Motion Capture Using Infrared Sensor Network
Interactive Full-Body Motion Capture Using Infrared Sensor Network
ijcga
 
PPT s01-machine vision-s2
PPT s01-machine vision-s2PPT s01-machine vision-s2
PPT s01-machine vision-s2
Binus Online Learning
 

What's hot (20)

Computer vision
Computer visionComputer vision
Computer vision
 
Open CV - 電腦怎麼看世界
Open CV - 電腦怎麼看世界Open CV - 電腦怎麼看世界
Open CV - 電腦怎麼看世界
 
Face detection ppt
Face detection pptFace detection ppt
Face detection ppt
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
 
Digital Image Processing and Edge Detection
Digital Image Processing and Edge DetectionDigital Image Processing and Edge Detection
Digital Image Processing and Edge Detection
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
 
Computer vision suprim regmi
Computer vision suprim regmi Computer vision suprim regmi
Computer vision suprim regmi
 
Image Processing and Computer Vision
Image Processing and Computer VisionImage Processing and Computer Vision
Image Processing and Computer Vision
 
AI Computer vision
AI Computer visionAI Computer vision
AI Computer vision
 
Application of edge detection
Application of edge detectionApplication of edge detection
Application of edge detection
 
Computer vision for interactive computer graphics
Computer vision for interactive computer graphicsComputer vision for interactive computer graphics
Computer vision for interactive computer graphics
 
Ai lecture 03 computer vision
Ai lecture 03 computer visionAi lecture 03 computer vision
Ai lecture 03 computer vision
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
 
Computer Vision - Artificial Intelligence
Computer Vision - Artificial IntelligenceComputer Vision - Artificial Intelligence
Computer Vision - Artificial Intelligence
 
computer vision & Opencv intro
computer vision & Opencv intro computer vision & Opencv intro
computer vision & Opencv intro
 
A350111
A350111A350111
A350111
 
Computer vision and robotics
Computer vision and roboticsComputer vision and robotics
Computer vision and robotics
 
Interactive full body motion capture using infrared sensor network
Interactive full body motion capture using infrared sensor networkInteractive full body motion capture using infrared sensor network
Interactive full body motion capture using infrared sensor network
 
Interactive Full-Body Motion Capture Using Infrared Sensor Network
Interactive Full-Body Motion Capture Using Infrared Sensor Network  Interactive Full-Body Motion Capture Using Infrared Sensor Network
Interactive Full-Body Motion Capture Using Infrared Sensor Network
 
PPT s01-machine vision-s2
PPT s01-machine vision-s2PPT s01-machine vision-s2
PPT s01-machine vision-s2
 

Similar to Computer vision series

Image Restoration for 3D Computer Vision
Image Restoration for 3D Computer VisionImage Restoration for 3D Computer Vision
Image Restoration for 3D Computer Vision
PetteriTeikariPhD
 
Defending
DefendingDefending
Defending
Sarah AL-Hzamat
 
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- ITOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
Anish Acharya
 
Iw3515281533
Iw3515281533Iw3515281533
Iw3515281533
IJERA Editor
 
Deferred Pixel Shading on the PLAYSTATION®3
Deferred Pixel Shading on the PLAYSTATION®3Deferred Pixel Shading on the PLAYSTATION®3
Deferred Pixel Shading on the PLAYSTATION®3
Slide_N
 
Practical Digital Image Processing 4
Practical Digital Image Processing 4Practical Digital Image Processing 4
Practical Digital Image Processing 4
Aly Abdelkareem
 
06_features_slides.pdf
06_features_slides.pdf06_features_slides.pdf
06_features_slides.pdf
JanuarAdiPutra3
 
Object recognition
Object recognitionObject recognition
Object recognition
saniacorreya
 
Recent advances of AI for medical imaging : Engineering perspectives
Recent advances of AI for medical imaging : Engineering perspectivesRecent advances of AI for medical imaging : Engineering perspectives
Recent advances of AI for medical imaging : Engineering perspectives
Namkug Kim
 
Segmentation
SegmentationSegmentation
Segmentation
guest49d49
 
Super Resolution of Image
Super Resolution of ImageSuper Resolution of Image
Super Resolution of Image
Satheesh K
 
I010634450
I010634450I010634450
I010634450
IOSR Journals
 
Performance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
Performance of Efficient Closed-Form Solution to Comprehensive Frontier ExposurePerformance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
Performance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
iosrjce
 
An ensemble classification algorithm for hyperspectral images
An ensemble classification algorithm for hyperspectral imagesAn ensemble classification algorithm for hyperspectral images
An ensemble classification algorithm for hyperspectral images
sipij
 
Secure System based on Dynamic Features of IRIS Recognition
Secure System based on Dynamic Features of IRIS RecognitionSecure System based on Dynamic Features of IRIS Recognition
Secure System based on Dynamic Features of IRIS Recognition
ijsrd.com
 
Fcv bio cv_cottrell
Fcv bio cv_cottrellFcv bio cv_cottrell
Fcv bio cv_cottrell
zukun
 
Fcv bio cv_cottrell
Fcv bio cv_cottrellFcv bio cv_cottrell
Fcv bio cv_cottrell
zukun
 
Computer Vision descriptors
Computer Vision descriptorsComputer Vision descriptors
Computer Vision descriptors
Wael Badawy
 
Digital Image Processing
Digital Image ProcessingDigital Image Processing
Digital Image Processing
Azharo7
 
Effective Pixel Interpolation for Image Super Resolution
Effective Pixel Interpolation for Image Super ResolutionEffective Pixel Interpolation for Image Super Resolution
Effective Pixel Interpolation for Image Super Resolution
IOSR Journals
 

Similar to Computer vision series (20)

Image Restoration for 3D Computer Vision
Image Restoration for 3D Computer VisionImage Restoration for 3D Computer Vision
Image Restoration for 3D Computer Vision
 
Defending
DefendingDefending
Defending
 
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- ITOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
TOWARDS OPTIMALITY OF IMAGE SEGMENTATION PART- I
 
Iw3515281533
Iw3515281533Iw3515281533
Iw3515281533
 
Deferred Pixel Shading on the PLAYSTATION®3
Deferred Pixel Shading on the PLAYSTATION®3Deferred Pixel Shading on the PLAYSTATION®3
Deferred Pixel Shading on the PLAYSTATION®3
 
Practical Digital Image Processing 4
Practical Digital Image Processing 4Practical Digital Image Processing 4
Practical Digital Image Processing 4
 
06_features_slides.pdf
06_features_slides.pdf06_features_slides.pdf
06_features_slides.pdf
 
Object recognition
Object recognitionObject recognition
Object recognition
 
Recent advances of AI for medical imaging : Engineering perspectives
Recent advances of AI for medical imaging : Engineering perspectivesRecent advances of AI for medical imaging : Engineering perspectives
Recent advances of AI for medical imaging : Engineering perspectives
 
Segmentation
SegmentationSegmentation
Segmentation
 
Super Resolution of Image
Super Resolution of ImageSuper Resolution of Image
Super Resolution of Image
 
I010634450
I010634450I010634450
I010634450
 
Performance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
Performance of Efficient Closed-Form Solution to Comprehensive Frontier ExposurePerformance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
Performance of Efficient Closed-Form Solution to Comprehensive Frontier Exposure
 
An ensemble classification algorithm for hyperspectral images
An ensemble classification algorithm for hyperspectral imagesAn ensemble classification algorithm for hyperspectral images
An ensemble classification algorithm for hyperspectral images
 
Secure System based on Dynamic Features of IRIS Recognition
Secure System based on Dynamic Features of IRIS RecognitionSecure System based on Dynamic Features of IRIS Recognition
Secure System based on Dynamic Features of IRIS Recognition
 
Fcv bio cv_cottrell
Fcv bio cv_cottrellFcv bio cv_cottrell
Fcv bio cv_cottrell
 
Fcv bio cv_cottrell
Fcv bio cv_cottrellFcv bio cv_cottrell
Fcv bio cv_cottrell
 
Computer Vision descriptors
Computer Vision descriptorsComputer Vision descriptors
Computer Vision descriptors
 
Digital Image Processing
Digital Image ProcessingDigital Image Processing
Digital Image Processing
 
Effective Pixel Interpolation for Image Super Resolution
Effective Pixel Interpolation for Image Super ResolutionEffective Pixel Interpolation for Image Super Resolution
Effective Pixel Interpolation for Image Super Resolution
 

Recently uploaded

Casting-Defect-inSlab continuous casting.pdf
Casting-Defect-inSlab continuous casting.pdfCasting-Defect-inSlab continuous casting.pdf
Casting-Defect-inSlab continuous casting.pdf
zubairahmad848137
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
171ticu
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
shadow0702a
 
Introduction to AI Safety (public presentation).pptx
Introduction to AI Safety (public presentation).pptxIntroduction to AI Safety (public presentation).pptx
Introduction to AI Safety (public presentation).pptx
MiscAnnoy1
 
Generative AI leverages algorithms to create various forms of content
Generative AI leverages algorithms to create various forms of contentGenerative AI leverages algorithms to create various forms of content
Generative AI leverages algorithms to create various forms of content
Hitesh Mohapatra
 
ACEP Magazine edition 4th launched on 05.06.2024
ACEP Magazine edition 4th launched on 05.06.2024ACEP Magazine edition 4th launched on 05.06.2024
ACEP Magazine edition 4th launched on 05.06.2024
Rahul
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
171ticu
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
co23btech11018
 
ML Based Model for NIDS MSc Updated Presentation.v2.pptx
ML Based Model for NIDS MSc Updated Presentation.v2.pptxML Based Model for NIDS MSc Updated Presentation.v2.pptx
ML Based Model for NIDS MSc Updated Presentation.v2.pptx
JamalHussainArman
 
Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
IJECEIAES
 
International Conference on NLP, Artificial Intelligence, Machine Learning an...
International Conference on NLP, Artificial Intelligence, Machine Learning an...International Conference on NLP, Artificial Intelligence, Machine Learning an...
International Conference on NLP, Artificial Intelligence, Machine Learning an...
gerogepatton
 
gray level transformation unit 3(image processing))
gray level transformation unit 3(image processing))gray level transformation unit 3(image processing))
gray level transformation unit 3(image processing))
shivani5543
 
Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...
bijceesjournal
 
Transformers design and coooling methods
Transformers design and coooling methodsTransformers design and coooling methods
Transformers design and coooling methods
Roger Rozario
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
KrishnaveniKrishnara1
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
Mahmoud Morsy
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Sinan KOZAK
 
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
RamonNovais6
 
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECTCHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
jpsjournal1
 
Textile Chemical Processing and Dyeing.pdf
Textile Chemical Processing and Dyeing.pdfTextile Chemical Processing and Dyeing.pdf
Textile Chemical Processing and Dyeing.pdf
NazakatAliKhoso2
 

Recently uploaded (20)

Casting-Defect-inSlab continuous casting.pdf
Casting-Defect-inSlab continuous casting.pdfCasting-Defect-inSlab continuous casting.pdf
Casting-Defect-inSlab continuous casting.pdf
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
 
Introduction to AI Safety (public presentation).pptx
Introduction to AI Safety (public presentation).pptxIntroduction to AI Safety (public presentation).pptx
Introduction to AI Safety (public presentation).pptx
 
Generative AI leverages algorithms to create various forms of content
Generative AI leverages algorithms to create various forms of contentGenerative AI leverages algorithms to create various forms of content
Generative AI leverages algorithms to create various forms of content
 
ACEP Magazine edition 4th launched on 05.06.2024
ACEP Magazine edition 4th launched on 05.06.2024ACEP Magazine edition 4th launched on 05.06.2024
ACEP Magazine edition 4th launched on 05.06.2024
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
 
ML Based Model for NIDS MSc Updated Presentation.v2.pptx
ML Based Model for NIDS MSc Updated Presentation.v2.pptxML Based Model for NIDS MSc Updated Presentation.v2.pptx
ML Based Model for NIDS MSc Updated Presentation.v2.pptx
 
Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
 
International Conference on NLP, Artificial Intelligence, Machine Learning an...
International Conference on NLP, Artificial Intelligence, Machine Learning an...International Conference on NLP, Artificial Intelligence, Machine Learning an...
International Conference on NLP, Artificial Intelligence, Machine Learning an...
 
gray level transformation unit 3(image processing))
gray level transformation unit 3(image processing))gray level transformation unit 3(image processing))
gray level transformation unit 3(image processing))
 
Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...
 
Transformers design and coooling methods
Transformers design and coooling methodsTransformers design and coooling methods
Transformers design and coooling methods
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
 
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
 
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECTCHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECT
 
Textile Chemical Processing and Dyeing.pdf
Textile Chemical Processing and Dyeing.pdfTextile Chemical Processing and Dyeing.pdf
Textile Chemical Processing and Dyeing.pdf
 

Computer vision series

  • 1. Imaging and Vision Pathfinding Perry Lea ACM Distinguished Lectures
  • 2. Floating Point 2D Graphics 3D Graphics Vision Computational Photography Physics Kernel Floating Point Requirement X Color Space Conversion Fixed Point X Gaussian Blur Fixed Point X Sobel Edge Detection Fixed Point X Bilateral Filters Fixed Point X Bilinear Interpolation Fixed Point X Bicubic Interpolation Half or Single Precision X Image Signal Processor Fixed Point X X Exposure Compensation Single Precision X X Image Blending Fixed Point X X Scaling Fixed (for binary scaling) X Texture Mapping Fixed Point X Pixel Shading Single / Double Precision X Z-Buffer Depth Test Single X Compositing Fixed Point or Half Precision X Ray Tracing Single Precision X 3D Vertex Shading Double Precision X Fluid Dynamics Single / Double Precision X JPEG Compression Fixed Point2
  • 4. Vision Segments 4 ADAS and Automotive Medical Imaging Consumer Electronics and GamingIndustrial Automation & Robotics Security, Surveillance, Intelligence, Defense Facial Recognition
  • 6. Vision Market 6 Tractica Research: 42% CAGR, $33B market by 2019 Market to Market: 22.6% CAGR, $22.2B market by 2020
  • 7. Human Vision | February 11, 2018 | Micron Confidential 7
  • 8. What do you see here? Do you see lines between the circles?  Guess what: there are none. Rule 1: Sensory input does not contain enough information to explain our perception What did you just see?  Did you see the people on the bridge?  Did you see the church?  Did you see the tunnel? Rule 2: There is too much sensory input to include in our coherent perceptions at any single moment 8
  • 9. Human Visual Dataflow Human vision interprets images bottom up and top down: Bottom Up: Based on raw sensory data (pixels) Top down: based on feature extraction Find the Target 9 Human Brain Visual System from Ganglion to Cortex
  • 10. How Human Vision Works Humans are born with a nearly fully developed vision system Cortical pathways are reinforced and restructured within the 1st year of development. Vision starts at ganglion cells and follows the optical nerve. Some receptors will excite with light intensity, some will inhibit activity. 1 0
  • 11. Feature Extraction When a collection of photoreceptors are organized into a center-surround field, the brain can easily perceive light and dark regions. Edges force ganglion cells to deliver reinforced or diminished signals. Visual System does an extraordinary job at throwing away information. 1 1 Ganglion Cell Signal Strength
  • 13. Vision Principles SIFT in 6 slides! Just as the human brain perceives image data top-down and bottom-up, so are typical vision algorithms.  Features are “interesting” parts of an image and we will rely on the same edges, corners, and ridges. To be useful, feature points must: Be numerous Be repeatable Represent orientation and scale Be fast to extract and match 1 3
  • 14. Typical Feature Extraction Algorithm Detector  Find Scale Space Extrema  Keypoint Localization Improve keypoints and throw out bad ones. Descriptor  Orientation Assignment (remove effects of rotation and scale)  Create Descriptor Use histograms of orientations 1 4 Lens Lens Correction White Balance Noise Reduction Demosaic Color Correction Tone Mapping Sharpening Gamma Correction 3A Stats RGB2YUV Scaler DRAM Image Signal Processor (Front End) Feature Extraction (Back End) 12 MegaPixel Image (RAW10=15 MB to 37MB. @30 fps = 450 MB/s) Preprocess Scan Image Filter Feature Locations Generate Signature Post Process Descriptors
  • 15. Finding Scale Space Finds keypoints in image. Image is convolved at different scales (variant of blob detection) Best way to do this is a Laplacian of Gaussian:  But a LoG is really computationally expensive (hmmm)  So we’ll cheat and do a Difference of Gaussian Blurs:  Convolved images are grouped by “octaves” which is simply the scale at that point. We convolve a certain number of images per octave k  Take the difference of the convolved images k per octave. 1 5
  • 16. Finding Scale Space Find Extreme Choose all extrema within a 3x3x3 neighborhood  This is done by comparing each pixel in the DoG images to its eight neighbors at the same scale and nine corresponding neighboring pixels in each of the neighboring scales. If the pixel value is the maximum or minimum among all compared pixels, it is selected as a candidate keypoint. 1 6
  • 17. Keypoint Localization Scale space extrema produce too many candidates. Minimize:  Use Taylor series expansion to get true extrema Reject:  Points with bad contrast  Points with strong edge response in 1 direction 1 7
  • 18. Orientation Assignment Remove effects of rotation Create a gradient of histograms (36 bins) Weighted by magnitude of Gaussian Window Any peak within 80% of highest is a new keypoint Parabola a parabola is fit to the 3 histograms closest to each peak 1 8
  • 19. Keypoint Descriptor We now want to compute a descriptor for each keypoint to make them distinctive with various illuminations, 3D views, etc. Similar to human biological vision Neurons respond to gradients at certain frequencies 4x4 gradient window with a histogram of 4x4 samples per window = 4x4x8 = 128 feature vectors 1 9 Lighting gains will not affect descriptors
  • 20. Feature Detection Algorithms Edge Detection: Canny, Sobel, Prewitt, Differential Corner Detectors: Harris, FAST, SUSAN Blob Detectors: Laplacian of Gaussian Difference of Gaussian Determinant of Gaussian 2 0  Transforms: – Ridge, Hough, Structural Tensor  Affine Invariants – Affine shape adapter – Harris Affine – Hessian Affine  Feature Descriptors – SIFT, SURF, GLOH, HOG, BRIEF, ORB, BRISK, FREAK
  • 21. Other Vision Challenges Segmentation Meaningful partitioning of image/video into non-overlapping regions and subvolumes. Ability to handle multi-modal data of varying complexity 2 1 Color Image Segmentation Output Original Image courtesy of University of California at Berkeley Courtesy RIT
  • 22. Other Vision Challenges Super Resolution Utilizing multiple images of a given scene to obtain a high resolution image with improved image quality 2 2
  • 23. Other Vision Challenges Hierarchical Scale Space Using information at various scales to determine the semantic structure of an image. Utilize probabilistic modeling of an image content to build a dynamic hierarchical tree for high resolution remote sensing. 2 3 Courtesy RIT
  • 24. Other Vision Challenges Computational Photography 2 4 Computational photography combines plentiful computing, digital sensors, modern optics, actuators, and smart lights to escape the limitations of traditional film cameras and enables novel imaging applications. Unbounded dynamic range, variable focus, resolution, and depth of field, hints about shape, reflectance, and lighting, and new interactive forms of photos that are partly snapshots and partly videos are just some of the new applications found in Computational Photography. • Light Field Arrays • Massive Image Stitching/Warping • Computational Optics • Holographic Imaging