Title : Image Signal Processing
Course No : EE5175
Credits : 4
Prerequisite :


Basics: Applications of image processing. Notion of pixel, resolution, quantization, photon noise.  Geometric transformations, source-to-target and target-to-source mapping, planar homography, rotational homography, change detection and mosaicing.

Image Formation:  Pin-hole versus real aperture lens model, lens as a 2D LSI system, blur circle, Doubly block circulant system matrix, pill box and Gaussian blur models, space invariant and space variant blurring. 3D Shape from Focus:  Depth of field, focal stack, focus operators, focus measure curve, Gaussian interpolation, 3D recovery, focused image recovery.Image Transforms: 1D Orthogonal transforms, Kronecker product, 2D orthogonal transforms from 1D, 2D DFT, 2D DFT for image matching, 2D DCT, Hadamard, KLT/PCA, PCA for face recognition, SVD, image denoising using SVD.

Image Enhancement:  Thresholding methods (peak-valley, Otsu, Chow-Kaneko), histogram equalisation and modification, Noise models, mean, weighted mean, median, weighted median, non local means filter, frequency domain filtering, illumination compensation by homomorphic filtering, segmentation by k-means clustering, higher-order statistics based clustering.

Image Restoration: Well-posed and ill-posed problems, Fredholm integral equation, condition number of  matrix, conditional mean, Inverse filter, Wiener filter, ML and MAP restoration, image super-resolution.

Edge Detection: Gradient operators, Prewitt, Sobel, Roberts, compass operators, LOG, DOG, Canny edge  detectors, non-maxima suppression, hysteresis thresholding.

Text Books :

  1. Gonzalez and Woods: Digital image processing.
  2. A.K. Jain: Fundamentals of digital image processing.

References : 

  1. Al Bovik: Handbook of image and video processing.
  2. J.S. Lim: Two dimensional signal and image processing.