1) Computer Vision
Linda G. Shapiro, George C. Stockman
2) Digital Image Processing (III Ed.) R. Gonzalez, R.E. Woods.
Pearson International Edition.
3) Course slides available from the Course Website
Learning Objectives
The course aims at providing the knowledge and ability to design and implement modules for the analysis and processing of image and video data.
- Design and prototype of processing modules for anayzing image regions based on color and texture- Design and prototype of processing modules for image segmentation
- Design and prototype of processing modules for detection and tracking of moving objects in video
Local operators. Linear and non-linear smoothing: Mean, Gaussian, bilateral filter, median; Correlation and template matching; Edge detection. First order operators: roberts, prewit, sobel. Second order operators: Laplacian, LoG, DoG. Canny edge detection. Parametric models for edge detection; Edge detection in color images.
Corner detection: Harris
Scale-space, geometric invariants and anisotropic diffusion models.
Color invariants.
Texture description: the co-occurrence matrix, the covariance matrix, Tamura features, Local Binary Patterns, Fourier Power Spectrum, Region covariance, Gabor, Moments;
Image thresholding: the Otzu model; local image thresholding;
Binary and lattice mathematical morphology: erode and dilate; open, close, binary reconstruction, distance transform, top-hat, bottom-hat, contrast enhancement, grayscale reconstruction, Granulometry with MM;
Image segmentation models. Hierarchical clustering, K-means, Gaussian Mixtures; Mean Shift; Normalized Cuts; Watersheds; Active contours models.
The Hough transform.
Video analysis: Motion field and optic flow. Optic flow estimation: Horn&Schunk, Lucas&Kanade, Polynomial expansion.