Computer vision

Computer Vision
BY : Mahmoud Hussein

What is Computer Vision?
 Make computers understand images and videos
 What kind of scene?
 Where are the cars ?
 How far is the building?

What is Computer Vision?
 Make computers understand images and videos
 Why is this happening?
 What is important ?
 What will I see?

Visual data on the Internet
 Flickr
 10+ billion photographs
 60 million images uploaded a month
 Facebook
 250+ billion
 300 million a day
 Instagram
 55 million a day
 YouTube
 100 hours uploaded every minute

Too big for humans
Need automatic tools to access and analyze visual
data!

Vision is Really Hard
 Vision is an amazing feature of natural intelligence
 Visual cortex occupies about 50% of Macaque brain
 More human brain devoted to vision than anything else
Is that a
queen or a
bishop

What did you see?
 Where this picture was taken?
 How many people are there?
 What are they doing?
 What object the person on the left standing on?
 Why this is a funny picture?

Computer: okay, it’s a funny picture

Computer Vision Matters
safety Health security
comfort fun Access

History of Computer Vision
 “In 1966, Minsky hired a first-year, undergraduate student and assigned him a
problem to solve over the summer:
 Connect a camera to a computer and get the machine to describe what it sees.
Marvin
Minsky, MIT
Turing award,
1969

1960’s: interpretation of synthetic worlds

1970’s: some progress on interpreting
selected images

1980’s: ANNs come and go; shift toward
geometry
and increased mathematical rigor

1990’s: face recognition; statistical analysis
in vogue

2000’s: broader recognition; large annotated
datasets available; video processing starts

2010’s: resurgence of deep learning
[AlexNet NIPS
2012]
[DeepFace
CVPR 2014]
[DeepPose
CVPR 2014]
[Show, Attend and Tell
ICML 2015]

Examples of Computer Vision Applications
 How is computer vision used today?
 Face detection
 Most digital cameras and smart phones detect faces (and more)
 For smart focus, exposure compensation, and cropping

Optical Character Recognition (OCR)
 Technology to convert scanned docs to text
 If you have a scanner, it probably came with OCR software

Matching street clothing photos in online Shops

Microsoft PhotoSynth: Photo Tourism

Computer vision for healthcare

Fundamentals of Computer Vision
 Light
 What an image records
 Matching
 How to measure the similarity of two regions
 Alignment
 How to align points
 Geometry
 How to relate world coordinates and image coordinates
 Categorization
 What similarities are important?

Image filtering
 Linear filtering: function is a weighted sum/difference of pixel values
 Really important!
 Enhance images
 Denoise, smooth, increase contrast, etc.
 Extract information from images
 Texture, edges, distinctive points, etc.
 Detect patterns
 Template matching

What is OpenCV?
 Open source Computer Vision library
 Cross-platform
 C/C++ API
 Originally developed by Intel
 Optimized and intended for real-time applications
 More than 500 functions available

Uses
 Human-Computer Interaction (HCI)
 Object Identification
 Object Recognition
 Face Recognition
 Motion Tracking
 Image Processing
 Mobile Robotics
 … and so on.

OpenCV modules (I)
 Core
 basic structures and algorithms
 Imgproc
 Image Processing algorithms
 Highgui
 built-in simple UI
 Video
 video analysis

OpenCV modules (II)
 Calib3d
 camera calibration and 3D reconstruction
 Features2d
 2D features framework
 Object detection
 detection of objects and other items

OpenCV modules (III)
 Ml
 machine learning classes used for statistical classification, regression and clustering
of data
 Gpu
 GPU-accelerated algorithms

Computer vision

More Related Content

What's hot (20)

Similar to Computer vision (20)

More from Mahmoud Hussein (20)

Recently uploaded (20)

Computer vision