Paper-review-YOLO-V4

- Paper : [Yolov4: Optimal speed and accuracy of object detection]

- Author : Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M.

- Cite as : arXiv preprint arXiv:2004.10934. (2020).

- YOLO-V4 structure diagram drawn by HeeSungP -

■ HeeSungP's presentation material (pdf, Korean)

> YOLO-V4_HeeSungP-SSL.pdf

■ Information

- YOLO-V4 Paper : https://arxiv.org/abs/2004.10934

- Author YOLO-V4 github : https://github.com/AlexeyAB/darknet

- Custom dataset YOLO-V4 implementation in Colab (train, test, evaluation) : Link

-> made by HeeSungP with author's C code

* This is the presentation materials in the Smart System Lab Seminar of Yonsei University. (2020-04-23)

* FIgures contained in the content were drawn directly or used in the reference below.

■

■ Reference

- YOLOv4: Optimal Speed and Accuracy of Object Detection

-> https://arxiv.org/abs/2004.10934

- CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features

-> https://arxiv.org/abs/1905.04899

- Cross-Iteration Batch Normalization

-> https://arxiv.org/abs/2002.05712

- Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition

-> https://ieeexplore.ieee.org/abstract/document/7005506?casa_token=uaAbG2H7U1UAAAAA:ZFraVsf7xy-cY4A1Nv3oBrENzdJ0AAhNjwqgF98tFGk67zusTbAon1BdPqTs4rkiLrTK6LR5gA