Topics

One page per concept: the reads, papers and projects about it in the order to read them, and the interactive figures pulled from those pages. The map at Explore shows how they connect.

Tasks

• Object detection
  30 pages · 41 figures
• Anchor-free detection
  1 page · 4 figures
• Segmentation
  4 pages · 1 figure
• Image classification
  8 pages · 4 figures
• Keypoints and pose
  2 pages · 4 figures
• Re-identification
  5 pages
• Tracking
  4 pages
• Anomaly detection
  2 pages
• Traffic analytics
  5 pages
• Saliency
  1 page
• Natural language
  10 pages
• Evaluation and metrics
  2 pages
• Video
  2 pages
• Faces
  3 pages
• Generative models
  2 pages
• Language models
  7 pages
• Vision and language
  3 pages · 6 figures
• Speech and audio
  1 page
• Medical imaging
  1 page
• Image enhancement
  1 page

Mechanisms

• Attention modules
  5 pages · 14 figures
• Self-attention
  4 pages · 9 figures
• Transformers
  8 pages · 7 figures
• Vision transformers
  6 pages · 11 figures
• Positional encoding
  2 pages · 4 figures
• Convolutions
  4 pages · 7 figures
• Backbone design
  8 pages · 12 figures
• Feature pyramids and necks
  6 pages · 11 figures
• Loss functions
  6 pages · 8 figures
• Class imbalance
  2 pages · 6 figures
• Non-maximum suppression
  2 pages · 6 figures
• Set prediction
  4 pages · 12 figures
• Activation functions
  1 page · 5 figures

Geometry

• Camera models
  5 pages · 15 figures
• Projective geometry
  5 pages · 14 figures
• Camera calibration
  3 pages · 9 figures
• Two-view geometry
  4 pages · 9 figures
• 3-D reconstruction
  4 pages · 9 figures
• Stereo vision
  2 pages · 4 figures

Training

• Self-supervised learning
  3 pages · 3 figures
• Semi-supervised learning
  4 pages
• Domain adaptation
  3 pages
• Knowledge distillation
  3 pages · 10 figures
• Data augmentation
  7 pages · 2 figures
• Training recipes
  10 pages · 13 figures
• Contrastive learning
  2 pages · 6 figures
• Pretraining at scale
  2 pages · 4 figures
• Weak supervision
  1 page
• Datasets and labelling
  3 pages

Efficiency

• Efficient attention
  5 pages · 13 figures
• Efficient networks
  4 pages · 1 figure
• Quantization
  1 page
• Pruning
  1 page
• Edge deployment
  4 pages · 3 figures
• Systems programming
  1 page
• Developer tools
  7 pages