Computer Vision

Master the full computer vision pipeline: image processing, CNNs, object detection, segmentation, vision transformers, and CLIP. For engineers who want to build production vision systems, not just use pre-trained models.

The route

  1. CNNs from Scratch — Deep Learning from Scratch (phase dl-01). Implement convolutions, pooling, and backprop through convolution layers. Build a CNN that classifies images — understanding exactly what every layer does.
  2. Computer Vision Engineering (full course) — Computer Vision Engineering. 6 phases, 28 lessons: image fundamentals, convolutions and CNNs, object detection (YOLO), segmentation (U-Net, Mask R-CNN, SAM), and modern vision (ViT, CLIP, DINO, diffusion). A capstone lesson builds a complete vision pipeline.
  3. CLIP & Multimodal Alignment — Multimodal AI Systems (phase mm-00). Implement CLIP from scratch and understand contrastive learning — the foundation of modern multimodal AI that aligns vision and language.

What you build

Convolutions and CNNs you implemented yourself, a complete vision pipeline in the course capstone, and CLIP built from scratch.

Every learning path

Create a free account — the opening phases of 29 of 35 courses are free, no credit card. Or see Pro pricing.

All courses · Pricing · About · FAQ · Glossary