Course

Introduction to Image Processing-Basic mathematical concepts: Image enhancement: Grey level transforms, Spatial filtering. Extraction of special features: edge and corner detection. Morphological processing, Image transforms, Discrete Fourier Transform, Fast Fourier Transform. Frequency domain enhancement.

Image Segmentation Algorithms: contextual, non-contextual segmentation, texture segmentation. Feature Detectors and Descriptors, Feature Matching-Object Recognition, Face detection (Viola Jones), Face Recognition,

Modern computer vision architectures based on deep convolutional neural networks, The Use of Motion in Segmentation Optical Flow & Tracking Algorithms, YOLO, DeepSORT: Deep Learning to Track Custom Objects in a Video, Action classification with convolutional neural networks, RNN, LSTM

Image registration, 2D and 3D feature-based alignment, Pose estimation, Geometric intrinsic calibration, -Camera Models and Calibration: Camera Projection Models – orthographic, affine, perspective, projective models. Projective Geometry, transformation of 2-d and 3-d, Internal Parameters, Lens Distortion Models, Calibration Methods – linear, direct, indirect and multiplane methods. Geometry of Multiple views- Stereopsis, Camera and Epipolar Geometry, Fundamental matrix; Homography, Rectification, DLT, RANSAC, 3-D reconstruction framework; Auto-calibration., Introduction to SLAM (Simultaneous Localization and Mapping).

Preamble

Computer Vision is one of the fastest growing and most exciting AI (Artificial Intelligence) disciplines in today’s academia and industry. This course is designed to open the doors for students who are interested in learning about the fundamental principles and important applications of computer vision. The course starts with the basic understanding of image formation and various image pre- processing techniques. It also deals with visual object detection and recognition algorithms. Object tracking and Motion segmentation from videos are also introduced as part of the course. The course also gives exposure to image reconstruction, camera calibration, stereo vision camera projection models etc. which enable students to understand the concepts required for implementing modern AI applications which can perceive understand and reconstruct complex visual world like robots navigating space and performing duties, smart cars which can drive safe etc.

Course Objectives

• To introduce students to the state-of-the-art algorithms in the area of image analysis and object recognition
• Give an exposure to video analysis techniques for object tracking and motion estimation
• To build good understanding on the computer vision concepts and techniques to be applied for robotic vision applications
• Enable students to apply the vision algorithms and develop applications in the domain of image analysis, robotic navigation

Course Outcomes

Prerequisites

• None.

CO-PO Mapping

Evaluation Pattern – 70:30

• Midterm Exam – 20%
• Lab Assignments – 25%
• Project – 25%
• End Semester Exam – 30%

Text Book / References

1. Deep Learning (Adaptive Computation and Machine Learning series) Ian Goodfellow, Yoshua Bengio, Aaron Courville, Francis Bach, January 2017, MIT Press
2. Richard Szelinski, Computer Vision: Algorithms and Applications, 2010
3. Trucco and A. Verri, Prentice Hall, 1998.Introductory techniques for 3D Computer Vision.
4. Marco Treiber, “An Introduction to Object Recognition Selected Algorithms for a Wide Variety of Applications”, Springer, 2010.
5. Forsyth and Ponce, “Computer Vision – A Modern Approach”, Second Edition, Prentice Hall, 2011.
6. C. Gonzalez, R. E. Woods, ‘Digital Image Processing’, 4th edition Addison-Wesley,2016.