Cascade Attentive Dropout for Weakly Supervised Object Detection
This work offers a strong specific gain in detection performance for researchers and practitioners working on weakly supervised object detection, an incremental improvement over existing methods.
This paper addresses the problem in weakly supervised object detection where models tend to focus only on the most discriminative parts of an object rather than the entire object. The authors propose a cascade attentive dropout strategy and an improved global context module, achieving 49.8% mAP and 66.0% CorLoc on PASCAL VOC 2007, outperforming existing state-of-the-art methods.
Weakly supervised object detection (WSOD) aims to classify and locate objects with only image-level supervision. Many WSOD approaches adopt multiple instance learning as the initial model, which is prone to converge to the most discriminative object regions while ignoring the whole object, and therefore reduce the model detection performance. In this paper, a novel cascade attentive dropout strategy is proposed to alleviate the part domination problem, together with an improved global context module. We purposely discard attentive elements in both channel and space dimensions, and capture the inter-pixel and inter-channel dependencies to induce the model to better understand the global context. Extensive experiments have been conducted on the challenging PASCAL VOC 2007 benchmarks, which achieve 49.8% mAP and 66.0% CorLoc, outperforming state-of-the-arts.