Computer Vision Roadmap

A practical path from zero to shipping CV. No filler.

0. Prerequisites

Linear algebra (just enough), Python, NumPy. Comfort with the command line.

1. Image fundamentals

Pixels, color spaces, convolution, filtering, edges, histograms.

2. Classical vision

Features (SIFT/ORB), homography, camera calibration, stereo, optical flow.

3. Geometry & 3D

Pinhole model, epipolar geometry, structured light, depth — see the DLP/structured-light gear for hands-on hardware.

4. Learning-based vision

CNNs, detection, segmentation, transformers for vision, self-supervision.

5. Practice

Reproduce a paper (see Notes). Ship one small thing end-to-end.