The definitive computer vision textbook, rebuilt chapter by chapter as interactive lessons. From image formation to neural rendering, with live simulations at every step.
What is computer vision? A brief history, the four approaches, and the grand challenges.
Geometric primitives, projective transforms, photometric image formation, the digital camera.
Point operators, linear filtering, convolution, Fourier transforms, pyramids, geometric warps.
Scattered data interpolation, variational methods, regularization, Markov random fields.
Supervised and unsupervised learning, deep neural networks, CNNs, advanced architectures.
Instance recognition, image classification, object detection, semantic segmentation, vision+language.
Points and patches, Harris corners, SIFT, edges and contours, lines, vanishing points, segmentation.
Pairwise alignment, RANSAC, image stitching, global alignment, blending and compositing.
Translational alignment, parametric motion, optical flow, Lucas-Kanade, Horn-Schunck, layered motion.
Photometric calibration, HDR imaging, super-resolution, image matting, texture synthesis.
Camera calibration, pose estimation, two-frame and multi-frame SfM, bundle adjustment, SLAM.
Epipolar geometry, stereo correspondence, local and global methods, multi-view stereo, deep networks.
Shape from X, active rangefinding, surface representations, volumetric methods, texture maps.
View interpolation, layered depth images, light fields, environment mattes, neural rendering.