Level 1 · Image Fundamentals & Classical CV

Before any neural network, an image is just a grid of numbers — and most of what "computer vision" actually does is linear algebra and statistics applied to that grid. This level builds the classical toolkit directly on real images with OpenCV: reading and manipulating pixels, filtering and convolution, thresholding, contour and shape analysis, keypoint detection and matching, and the classical (pre-deep-learning) detection techniques that still matter for simple, constrained problems.

By the end of this level you can read, transform, filter, segment, and analyze a real image using nothing but classical image processing — and you'll have built a real, working document-scanner pipeline from scratch.