End-to-End Multi-Task Learning with IIL for Edge and Segmentation Feature Maps
To facilitate efficient multi-task learning, we employ an end-to-end approach that integrates IIL to simultaneously supervise and learn the two branches. Specifically, we represent both the edge feature map and the segmented feature map using binary values.
原文地址: https://www.cveoy.top/t/topic/qB8g 著作权归作者所有。请勿转载和采集!