Unit 1 Digital Image Fundamentals and Acquisition

6 lecture hours

• – Origins and applications of digital image processing – Image acquisition: sensors
• sampling and quantization – Pixel intensity relationships: neighbors
• connectivity – Basic intensity transformations: negative, log, power-law (gamma)

Unit 2 Spatial and Frequency Domain Image Enhancement

6 lecture hours

• – Histogram processing: equalization
• local methods – Spatial filtering: smoothing kernels
• sharpening and Laplacian – Frequency-domain: 2-D DFT properties
• ideal and Butterworth LPF/HPF

Unit 3 Edge Detection, Morphology and Segmentation

8 lecture hours

• – Edge detection: Sobel
• Prewitt, LoG, Canny edge detector – Morphological processing: erosion
• closing – Morphological algorithms: hit-or-miss
• skeletonisation – Segmentation: region growing
• region splitting and merging

Unit 4 Image Feature Detection and Description

7 lecture hours

• – Texture descriptors: LBP, co-occurrence matrix (GLCM)
• Gabor – Interest point detection: Harris corners
• FAST – Feature descriptors: HOG, SIFT
• SURF overview

Unit 5 Camera Models and Projective Geometry

• – Projective geometry: homogeneous coordinates
• projective transforms – Camera model: intrinsic matrix
• extrinsic parameters

Unit 6 Deep Learning for Object Detection and Segmentation

6 lecture hours

• – CNN architectures: AlexNet, VGG, ResNet
• EfficientNet overview – Object detection: R-CNN family
• SSD – Semantic segmentation: FCN, U-Net
• DeepLab (atrous convolution)

Unit 7 Motion Analysis and Object Tracking

6 lecture hours

• – Optical flow: Horn-Schunck global method
• Lucas-Kanade local method – Motion segmentation: background subtraction, GMM, frame differencing – Object tracking: Kalman filter
• particle filter
• mean-shift tracker