论文地址: Slicing Aided Hyper Inference and Fine-tuning for Small Object Detection开源代码: https://github.com/obss/sahi摘要场景中的小物体检测是视频监控的主要挑战。这类物体由于图像中的像素占比较少…
2 Slicing Aided Hyper Inference (SAHI) 和滑动窗口的思想类似,作者将图像分割成多个重叠的切片slices。这样就使得输入网络的图像的小目标占比较高。 2-1 Slicing aided fine-tuning 从fine-tuning数据中,随机裁剪出一个随机区域([ , ]和[ , ]),等比例放缩到输入图像大小,和原图一起用于fine-tuning。 注意:...
In this work, an open-source framework called Slicing Aided Hyper Inference (SAHI) is proposed that provides a generic slicing aided inference and fine-tuning pipeline for small object detection. The proposed technique is generic in the sense that it can be applied on top of any available ...
Slicing Aided Hyper Inference and Fine-tuning for Small Object Detection. arXiv 2022, arXiv:2202.06934. [Google Scholar] Zhang, H.; Hao, C.; Song, W.; Jiang, B.; Li, B. Adaptive Slicing-Aided Hyper Inference for Small Object Detection in High-Resolution Remote Sensing Images. Remote ...