Topics on Computer Vision

Course outline

A. Theory

• Computer vision, Image formation and optical sensors
• Elements from projective geometry, camera’s calibration, linear and nonlinear estimation of intrinsic and external camera’s parameters 
• Photogrammetry, shadows and colors 
• Parametric curves and surfaces 
• Multidimensional signals and systems. Multidimensional linear spatio-temporal systems, Multidimensional filters, filters Gabor, wavelets, Scale-space decomposition, pyramids
• Stereopsis and Multiview geometry, scene reconstruction using two and multiple images 
• Image matching and alignment, geometric and photometric distortions, Modeling geometric distortions via linear (affine) and nonlinear (homographies) transformations, mosaicking. 
• Feature based Image matching, detection and extraction of features, Features based on corners, blobs, SIFT, Laplacian, DoG and SURF detectors
• Super resolution 
• Motion and optical flow estimation, video stabilization. 
• Machine Learning, Neural Networks, Deep Neural Nets
• Oblect Detection, Classical and Deep techniques

B. Laboratory Exercises

• Exercise 1: Basic geometric transformations and their use in the animation 
• Exercise 2: Image Pyramids, Image de-noising and feature detection and extraction 
• Exercise 3: Scene reconstruction using a stereo image system 
• Exercise 4: Area based Image alignment and Joint alignment of  set of images
• Exercise 5: Feature based Image alignment
• Exercise 6-8: Open source learning platforms pytorch and tensorflow
• Exercise 9-10: Implementation of state of the art object detection techniques