Computer Vision — Roadmap

1. Vision Fundamentals

• Image representation: pixels, channels, color spaces
• Convolution as classical filtering (before it's a neural network layer)
• Morphological operations: erosion, dilation, opening, closing
• Edge detection: Sobel, Canny
• Classical feature extraction: SIFT/ORB, HOG
• Data augmentation for vision

2. Vision Tasks & Models

• Classification, detection, segmentation (recap — see Vision Architectures for the models)
• Pose estimation
• OCR (Optical Character Recognition)
• Object tracking
• Depth estimation
• Optical flow

3. Modern Vision & Multimodal

• Vision Transformers in production (see Vision Architectures)
• Vision-language models and image-text alignment (CLIP)
• Image generation (see Generative Models)
• Video understanding
• Video generation
• 3D vision: point clouds, NeRF, 3D reconstruction

Next: NLP — the equivalent classical-to-modern arc for text.