Digital Image Processing — Gonzalez & Woods

Rafael C. Gonzalez and Richard E. Woods, 4th Edition (Pearson)

The standard text on digital image processing. It carries the one-dimensional signal processing of the earlier phases into two dimensions: the image as a sampled, quantized 2-D signal, then spatial filtering (which is 2-D convolution), frequency-domain filtering (the 2-D DFT), and image restoration (the Wiener filter, applied to a blurred and noisy image). On that base it builds the algorithms that only exist in the image domain — compression and its entropy coders, segmentation, and feature extraction — and closes, in the 4th edition, on a substantial new chapter on neural networks and deep convolutional networks.

Deeper reading for Course 1 — Lesson 42 (spatial and frequency-domain filtering, restoration, edges, and features; Ch. 2–5, 8, 10–11), then Ch. 12 (image pattern classification and deep CNNs) behind the learned-DSP close of Lesson 43. Every chapter carries a problem set, worked by hand.

Chapters

Exercises added as I work through each chapter.