Course

Syllabus

Introduction to Computer Vision Image Formation and Representation Color Models Camera Calibration and Lens Distortion Basics of Embedded Architectures for Vision ARM Cortex, Raspberry Pi, NVIDIA Jetson, FPGAs Interfacing Cameras with Embedded Platforms.Image Enhancement and Filtering Edge Detection (Sobel, Canny) Morphological Operations Feature Detection (Harris, FAST, ORB) Object Recognition using Feature Descriptors Embedded Optimization for Preprocessing Real-Time Constraints and Memory Considerations.Face and Object Detection using Haar Cascades, HOG, and CNNs Motion Detection and Tracking (Kalman Filter, Optical Flow) Lightweight Deep Learning Models (MobileNet, YOLO-tiny) Model Deployment with Tensor Flow Lite and OpenCV on Edge Devices Case Studies: Smart Surveillance, Vision for Robotics, IoT Cameras.

Text Books / References

1. Simon J. D. Prince, “Computer Vision: Models, Learning, and Inference”, Cambridge University Press, First Edition, 2012.
2. Szeliski Richard, “Computer Vision: Algorithms and Applications”, Springer, Second Edition, 2022.
3. Adrian Kaehler and Gary Bradski, “Learning OpenCV 4: Computer Vision with Python”, O’Reilly Media, Second Edition, 2019.
4. Joseph Howse, “OpenCV for Embedded Systems and IoT”, Packt Publishing, First Edition, 2020.
5. David G. Lowe, “Embedded Computer Vision”, Springer, First Edition, 2014.

Pre-requisite: Nil

Course Objectives:

• To introduce the fundamentals of computer vision and image processing with a focus on embedded system constraints.
• To equip students with practical skills to implement vision algorithms for object detection, recognition, and tracking.
• To familiarize students with hardware platforms and optimization techniques for deploying vision systems in real-time.
• To enable students to design efficient computer vision applications for domains such as robotics, surveillance, and IoT using embedded devices

Course Outcomes:

• CO1: Understand the principles of computer vision and their relevance to embedded platforms.
• CO2: Apply image preprocessing, feature extraction, and object recognition techniques.
• CO3: Implement real-time computer vision algorithms on embedded systems.
• CO4: Evaluate and optimize performance of embedded vision applications for power, speed, and accuracy.

 CO-PO Mapping: