PAHSU · Final Year B.Tech. Electronics and Telecommunication Engineering · NEP 2020
Computer Vision
ENTPEC-04B · Programme Elective IV · Semester VII · 4 credits · End-semester 100 only
One of the Programme Elective IV options. Programme Elective IV can instead be an NPTEL/SWAYAM MOOC from the Board of Studies' list.
Unit 1: Essential Mathematical Tools
8 hoursLeast squares, probabilistic inference, robust optimisation beyond RANSAC, eigen-analysis, PCA, SVD, clustering and gradient-based optimisation; camera models, epipolar geometry, stratified reconstruction and large-scale reconstruction.
Examined in the 100-mark end-semester paper
- RANSAC and line fitting
- PCA and LDA
- Singular value decomposition
- Camera models and calibration
- Epipolar geometry and stereo
- Structure from motion and bundle adjustment
Unit 2: Feature Detection and Tracking
8 hoursThe Harris corner detector, KLT tracking, SIFT and other contemporary descriptors; learned feature descriptors and deep feature matching.
Examined in the 100-mark end-semester paper
- Corner detection
- SIFT, SURF and BRISK features
- Feature detectors and descriptors
- Feature matching and tracking
Unit 3: Segmentation
8 hoursLow-level segmentation by energy minimisation and clustering; semantic, instance and panoptic segmentation; vision transformers for segmentation.
Examined in the 100-mark end-semester paper
- Graph-based segmentation: normalised cuts and mean shift
- Semantic segmentation with U-Net
- Instance and panoptic segmentation
- Vision transformers
Unit 4: High-Level Vision
8 hoursCNNs, single-image depth estimation and FlowNet; 3D object recognition and reconstruction; motion: triangulation, bundle adjustment, translational and parametric alignment, spline-based and layered motion, optical flow.
Examined in the 100-mark end-semester paper
- CNN architectures: LeNet, AlexNet and VGG
- Stereo vision and depth
- 3D object recognition and reconstruction
- Optical flow and layered motion
Unit 5: Synthesis
6 hoursGANs, 3D shape synthesis, combining viewpoint and texture, semantic image synthesis; diffusion models, text-to-image and multimodal generation, 3D generative models.
Examined in the 100-mark end-semester paper
- Generative adversarial networks
- Diffusion models and text-to-image generation
Unit 6: Applications
10 hoursSurveillance: foreground-background separation, particle filters, Chamfer matching, tracking and occlusion, combining views from several cameras, human gait analysis; in-vehicle vision: locating the roadway and markings, road signs and pedestrians.
Examined in the 100-mark end-semester paper
- Background subtraction and motion segmentation
- Surveillance, particle filters and tracking
- In-vehicle vision: road, sign and pedestrian detection
Unit contents are summarised from the official PAHSU syllabus . AiCanCode.org is not affiliated with Punyashlok Ahilyadevi Holkar Solapur University (PAHSU).