This course offers a comprehensive, hands-on introduction to the field of computer vision, blending foundational principles with modern deep learning techniques. Students will begin by exploring the basics of image formation, manipulation, and traditional image processing methods like filtering, edge detection, and feature description. The curriculum then progresses to advanced topics, including data preparation for vision tasks, image classification with Convolutional Neural Networks (CNNs), and state-of-the-art methods for object detection, tracking, and image segmentation. The final weeks will cover the exciting domain of generative models, such as Autoencoders and GANs, and overview of 3D vision.
The course is structured around weekly lectures and labs, providing a strong practical component. Through a series of programming assignments, students will gain direct experience implementing and applying the algorithms discussed in class. Key assessments include a midterm exam, multiple assignments, and a final project, culminating in a presentation and final exam. The course is designed to equip students with a solid theoretical understanding and the practical skills necessary to solve real-world computer vision problems.
| Week | Date | Lecture (9 - 11 AM) | Lab (11 AM - 12PM, 1 - 3 PM) | Events and Deadlines |
|---|---|---|---|---|
| 1 | Aug 16, 2025 | Introduction to Computer Vision - History, applications, challenges. - Image formation, color spaces (RGB, HSV, Grayscale). - Point Operation - Course overview - Course logistic |
Lab 1: Your First Computer Vision Program - Read Image from file - Inspect Image Properties : shape, channels, dtype - Display the image - Convert to other color spaces - Change brightness, contrast (point operation) - Save image to disk - Image Manipulations: Cropping, drawing rectangle onto image |
|
| Aug 23, 2025 | NO Class | |||
| 2 | Aug 30, 2025 | Image Filtering and Enhancement - Linear filters - Non-Linear filters - Histograms & histogram equalization. |
Lab 2: Image Sculpting Implement and apply histogram-based enhancement, linear, and non-linear filtering to modify and analyze images. - Histogram Analysis and Equalization - Convolution from Scratch - Filtering with OpenCV's Optimized Functions |
|
| 3 | Sep 6, 2025 | Feature Detection - Edge detection (Canny). - Corner detection (Harris). - Line Detection (Hough Transform) |
Lab 3: Finding Primitives - Harris Corner Detection - Canny Edge Detection - Hough Line Transform |
|
| 4 | Sep 13, 2025 | Feature Description & Matching - SIFT - Feature Matching - Image Stitching - Panorama |
Lab 4: Describing and Matching Features Image Stitching Pipeline: - Detect keypoints and extract descriptors - Match features between the two images - Compute Homography with RANSAC - Warp image A to align with image B's perspective - Stitch the images together |
|
| 5 | Sep 20, 2025 | Camera Calibration & Geometry - Pinhole camera model - intrinsic/extrinsic parameters - camera calibration. |
Lab 5: Correcting a Distorted View 'undo' the distortion, transforming flawed real-world images into the perfect, undistorted images predicted by the pinhole model. - Camera Calibration - Applying the Calibration - Undistortion - (Bonus) Augmented Reality |
|
| 6 | Sep 27, 2025 | Image Classification - kNN - Linear Classifiers |
Lab 6: Introduction to Classification - kNN - Linear Classifier Data Label Tool (CVAT, LabelImg, Label Studio) - Labeling for Image Classification - Labeling for Object Detection - Labeling for Segmentation |
Assignment #1 Due Assignment #2 Release |
| 7 | Oct 4, 2025 | Midterm Exam | ||
| 8 | Oct 11, 2025 | Image Classification with CNNs - Optimization - Neural Network basics : MLP, backprop - CNN - CNN architectures: LeNet, AlexNet, GoogLeNet, VGG, ResNet |
Lab 7: Your First CNN Classifier - MLP - CNN - ResNet15 |
|
| 9 | Oct 18, 2025 | Object Detection - Two-stage detectors: R-CNN family - Single-stage detectors - YOLO |
Lab 8: You Only Look Once (YOLO) - Yolo |
Assignment #2 Due |
| 10 | Oct 25, 2025 | Object Tracking Classical filtering-based trackers - Kalman Filter - MeanShift |
Lab 9: Follow That Object - Kalman Filter - MeanShift |
|
| 11 | Nov 1, 2025 | Image Segmentation - Traditional Image Segmentation Techniques: Thresholding, Region-based, Clustering, MeanShift - Learning-based Image Segmentation: FCN, U-Net, Mask R-CNN, SAM |
Lab 10: Pixel-Perfect Predictions - UNet - SAM |
Assignment #3 Release |
| 12 | Nov 8, 2025 | Generative Models - Autoencoder - Variational Autoencoder - Genrative Adversarial Network, DCGAN |
Lab 11: Generating New Realities - AE, VAE - GAN |
|
| 13 | Nov 15, 2025 | Overview of 3D Vision - 3D shape Representations - 3D Deep Learning - Multiview images - Point Cloud - Voxel - Polygon Mesh |
Lab 12: Point Cloud Classification - From Mesh to Point cloud, and Voxel - PointNet |
Assignment #3 Due |
| 14 | Nov 22, 2025 | Project Presentation | Project Report Due | |
| 15 | Nov 29, 2025 | Final Exam |