P stage 3 - 한사랑 개발회

참고자료

Object Detection

mmdetection관련 자료

• mmdetection 자료 : https://github.com/open-mmlab/mmdetection
• mmdetection documentation : https://mmdetection.readthedocs.io/en/latest/
• efficientnet & mmdetection : https://github.com/HardLaugh/EfficientDet-bifpn
• swin Transform & mmdetection : https://github.com/SwinTransformer/Swin-Transformer-Object-Detection

code library github

- mmcv : https://github.com/open-mmlab/mmcv

Model 관련 논문

two stage

- rcnn : https://arxiv.org/pdf/1311.2524.pdf
- fast rcnn : https://arxiv.org/pdf/1504.08083.pdf
- faster rcnn : https://arxiv.org/pdf/1506.01497.pdf
- mask rcnn : https://arxiv.org/pdf/1703.06870.pdf

multi stage

- cascade mask rcnn : https://arxiv.org/pdf/1712.00726v1.pdf
- hybrid task cascade : https://arxiv.org/pdf/1901.07518v2.pdf
- Backbone
- AN IMAGE IS WORTH 16X16 WORDS: https://arxiv.org/pdf/2010.11929.pdf
- DeiT : https://arxiv.org/pdf/2012.12877.pdf
- Swin Transformer : https://arxiv.org/pdf/2103.14030.pdf

Sidedish

- FPN : https://arxiv.org/pdf/1612.03144.pdf
- PAN : https://arxiv.org/pdf/1803.01534.pdf