1. YOLOv4: Optimal speed and accuracy of object detection;bochkovskiy;arXiv 2004 10934,2020
2. SIoU loss: More powerful learning for bounding box regression;gevorgyan;arXiv 2205 12740,2022
3. YOLOv3: An incremental improvement;redmon;arXiv 1804 02767,2018
4. SimAM: A simple, parameter-free attention module for convolutional neural networks;yang;Proc 38th Int Conf Mach Learn,2021
5. X-SDD: A New Benchmark for Hot Rolled Steel Strip Surface Defects Detection