Comprehensive Attention Self-Distillation for Weakly-Supervised Object Detection
This work addresses the challenge of training object detectors with only image-level labels, which is crucial for reducing annotation costs in computer vision, though it is incremental as it builds on existing WSOD methods.
The paper tackled the problem of weakly supervised object detection (WSOD) being prone to detecting only salient objects, clustered objects, or parts, and lacking consistency across image transformations, by proposing Comprehensive Attention Self-Distillation (CASD), which achieved new state-of-the-art results on PASCAL VOC 2007/2012 and MS-COCO benchmarks.
Weakly Supervised Object Detection (WSOD) has emerged as an effective tool to train object detectors using only the image-level category labels. However, without object-level labels, WSOD detectors are prone to detect bounding boxes on salient objects, clustered objects and discriminative object parts. Moreover, the image-level category labels do not enforce consistent object detection across different transformations of the same images. To address the above issues, we propose a Comprehensive Attention Self-Distillation (CASD) training approach for WSOD. To balance feature learning among all object instances, CASD computes the comprehensive attention aggregated from multiple transformations and feature layers of the same images. To enforce consistent spatial supervision on objects, CASD conducts self-distillation on the WSOD networks, such that the comprehensive attention is approximated simultaneously by multiple transformations and feature layers of the same images. CASD produces new state-of-the-art WSOD results on standard benchmarks such as PASCAL VOC 2007/2012 and MS-COCO.