Computer Vision Section 059
Object Detection in Depth
Every moving part of a detector, from box formats to transformers that skip anchors entirely.
13 of 13 lessons published Three reading levels on every lesson
Lessons in order
Work top to bottom. Each lesson assumes the one above it.
- Bounding box formats
- Intersection over union
- Anchor boxes
- Non-maximum suppression
- Feature pyramid networks
- RoI pooling and RoI align
- R-CNN, Fast R-CNN and Faster R-CNN
- YOLO versions compared
- Focal loss and RetinaNet
- Anchor-free detection
- DETR and set prediction
- Open-vocabulary detection
- Detecting very small objects