Motion Fields and Optical Flow
How camera motion and scene depth determine image velocity: the motion field equation, pure translation and rotation, the focus of expansion, time to contact, and how optical flow differs from the motion field.
Technical notes with derivations, implementation details, and visualizations. Geometry, estimation, robotics perception, and generative models.
Build your phone's panorama mode from scratch: Harris corners, ANMS, descriptors, matching, RANSAC, homographies, warping and blending - with interactive demos and runnable code at every step.
How camera motion and scene depth determine image velocity: the motion field equation, pure translation and rotation, the focus of expansion, time to contact, and how optical flow differs from the motion field.
A practical guide to authoring .mdx posts in this site: frontmatter, headings, math, images, code blocks, and interactive Python cells via PyodideRunner.
Formulate supervised classification as decision-making over discrete labels, contrasting geometric and probabilistic views, and introducing margin-based methods through Support Vector Machines
Fit and visualize linear, quadratic, cubic, and degree-5 polynomial regressions on mildly nonlinear data