💬Comments welcome. To leave a note, select any text and click the note / highlight button that pops up — or open the panel with the tab at the top-right (‹). Notes are visible only inside our private review group.
jump to

1.4 Problem sets

The surest test of understanding a method is to implement it. The book's ideas are meant to be built, and these problem sets are how: each takes one chapter's core algorithm and asks you to write it from scratch and run it on real photographs. They were developed for MIT's Digital & Computational Photography course, so their sequence and coverage do not track the book chapter-by-chapter — the assignments follow the course's own arc (point operations and color, then convolution and denoising, then the multi-image and geometry methods, then an open project), and each maps to one or more chapters rather than lining up one-to-one with them. The figures throughout the book that show a method's genuine output (a merged radiance map, a stitched panorama, a recovered optical-flow field) are produced by exactly these implementations.