Presented by
Dinesh Anand S.P

---

[attachment=13034]
 introduction
 Processing large set of image data until 1970’s was difficult.
 Still in the state of basic research.
 Computer vision is the science and technology of machines that see.
 Sub domains include event detection, video tracking, object recognition etc
Biological vision(a) Vs Computer vision(b).
Relation between computer vision and other fields
 Digital image processing
 Digital Image
 Representation of Digital image
 Digital Image Processing
 Image Processing (low level)
 Image Analysis
 Image Understanding
 Typical tasks of computer vision
 Recognition
 Object Recognition
 Identification
 Detection
 Motion analysis
 Ego motion
 Tracking
 Optical flow
 Scene Reconstruction
 Image restoration
 Computer vision systems
 Image acquisition
 Pre-processing
 Feature extraction
 Detection/segmentation
 High-level processing
 Introduction to object tracking
 Object tracking is an important task of Computer vision
 Tasks of object tracking includes
 Motion based recognition
 Automated surveillance etc
 Tracking can be complex because of
 Loss of information
 Complex object motion etc
 Object representation
 Points
 Primitive geometric shapes
 Articulated shaped models
 Object silhouette and contour
 Skeletal model
 Probability densities of object appearance
 Templates
 Active appearance models
 Multi view appearance models
 Contd..
Object representations
 Feature selection for tracking
 Color à
 Edges à
 Optical flow à
 Texture à
 Video play for color tracking
 Object Detection
 Point detectors
 Background subtraction
 Segmentation
 Supervised learning
 Object Tracking
 applications
 Controlling processes
 Detecting events
 Industrial inspection/quality control
 Recognitions & biometrics
 Real time environments etc
 In animation and cinematography
 Material inspection application
 Advantages, Limitations &
Future works
 Advantages
 Computer vision systems can work continuously.
 Motion camera
 Can track position and orientation of multiple objects
 Excellent flexibility and real time execution etc
 Limitations (drawbacks)
 Computer skill is needed and the systems are costly
 Sensitive to light and obstruction
 3-D problem called perspection and occlusion
 Contd..
 Future works
 Use of multiple cameras
 High speed units and graphics
 Complete vision processing system on a single chip
 Conclusion
 Computer vision is a vast field of vision science.
 It is very much useful in tracking application.
 It takes the aid of Image Processing to build computer intelligence.
 Applied in wide variety of fields like Machine vision, Artificial intelligence etc.
 Detailed summaries of Object tracking, including object representations, motion models etc.
