Semester 7Year 4 · OddCore Subject★★★★★ Hard
CS 704

Computer Vision

Study of image processing, CNNs, object detection, segmentation, face recognition, and visual AI applications.

4Units
28Topics
4Credits
60hLecture hrs
100Max marks
Your Progress
0 / 28 topics
0% complete
Overview
🎯
Why it matters
Self-driving cars see the road. Medical AI diagnoses from X-rays. Face unlock on phones. Augmented reality apps. Computer Vision is teaching machines to see and understand the visual world.
💼
Placement relevance
CV Engineer roles at Tesla, NVIDIA, autonomous vehicle companies. Medical imaging startups. AR/VR companies (Meta, Apple). ₹30-65 LPA for CV specialists.
🔗
Prerequisites for
Autonomous Vehicles · Medical Imaging · AR/VR Development · Robotics · Surveillance Systems · Visual AI Research
📚
Recommended books
Computer Vision: Algorithms and Applications by Richard Szeliski · Deep Learning for Computer Vision by Rajalingappaa Shanmugamani · Programming Computer Vision with Python by Jan Erik Solem · Multiple View Geometry in Computer Vision by Hartley and Zisserman
Curriculum — 4 Units
U1
Unit 1 · 7 Topics · 0% complete
Image Processing Fundamentals
⚡ Key Formulae
Convolution:G(x,y) = Σ Σ I(x+i, y+j) × K(i,j)
Sobel:Gx, Gy kernels for edge detection
Image Representation (RGB, Grayscale)
Filtering (Gaussian, Median)
Edge Detection (Sobel, Canny)
Morphological Operations
Histogram Equalization
Feature Extraction (SIFT, SURF, ORB)
Image Transformations
U2
Unit 2 · 7 Topics · 0% complete
Deep Learning for Vision
⚡ Key Formulae
IoU:Intersection over Union = Area(overlap) / Area(union)
Non-Max Suppression:Filter overlapping boxes (keep highest confidence)
CNNs for Image Classification
Transfer Learning (VGG, ResNet, Inception)
Data Augmentation Techniques
Object Detection (R-CNN, Fast R-CNN)
YOLO (You Only Look Once)
SSD (Single Shot Detector)
Bounding Box Regression
U3
Unit 3 · 7 Topics · 0% complete
Segmentation & Advanced Tasks
⚡ Key Formulae
Dice Coefficient:2×|A∩B| / (|A| + |B|) — segmentation metric
U-Net:Encoder-Decoder with skip connections
Semantic Segmentation
Instance Segmentation
U-Net Architecture
Mask R-CNN
Image Captioning
Visual Question Answering
Style Transfer
U4
Unit 4 · 7 Topics · 0% complete
Specialized Applications
⚡ Key Formulae
Face Recognition:Extract embeddings, compute similarity (cosine/Euclidean)
OCR Pipeline:Detection → Recognition → Post-processing
Face Detection & Recognition
Pose Estimation
Optical Character Recognition (OCR)
Image Generation (GANs, Diffusion)
3D Vision Basics
Video Analysis
Real-time Processing
Previous Year Questions
Unit 12023 · End Semester10 marks
Apply Canny edge detection on a given image. Explain the steps: Gaussian smoothing, gradient calculation, non-maximum suppression, hysteresis thresholding. Show intermediate results.
Unit 22023 · End Semester8 marks
Explain YOLO architecture for object detection. How does it differ from R-CNN? Calculate IoU for two bounding boxes: Box1(10,10,50,50), Box2(30,30,70,70).
Unit 32022 · End Semester6 marks
What is semantic segmentation? Explain U-Net architecture with a diagram. How is it different from object detection?
Exam Strategy
🖼️
Image processing algorithms
Canny edge detection, SIFT features, histogram equalization — know step-by-step procedures. Show intermediate images/matrices.
📦
Object detection formulas
IoU calculation, Non-Max Suppression, Anchor boxes. Practice bounding box math. Compare YOLO vs R-CNN (speed vs accuracy).
🎨
Draw CNN architectures
U-Net, ResNet, YOLO diagrams expected. Label conv layers, pooling, skip connections. Specify filter sizes and output dimensions.
Related Subjects
Semester 7
Deep Learning
CS 701
Semester 5
Computer Graphics
CS 504
Semester 5
Machine Learning
CS 501