Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation.
Zheng Z, Wang P, Ren D, Liu W, Ye R, Hu Q, Zuo W.
Zheng Z, et al.
IEEE Trans Cybern. 2022 Aug;52(8):8574-8586. doi: 10.1109/TCYB.2021.3095305. Epub 2022 Jul 19.
IEEE Trans Cybern. 2022.
PMID: 34437079
In particular, we consider three geometric factors, that is: 1) overlap area; 2) normalized central-point distance; and 3) aspect ratio, which are crucial for measuring bounding-box regression in object detection and instance segmentation. The three geometric factors are then inc …
In particular, we consider three geometric factors, that is: 1) overlap area; 2) normalized central-point distance; and 3) aspect ratio, whi …