Object Region Mining with Adversarial Erasing: A Simple Classification to Semantic Segmentation Approach

TLDR

Classification networks typically respond only to small, sparse discriminative regions, which is insufficient for the dense, pixel‑wise localization required by semantic segmentation. The study proposes an adversarial erasing method that progressively mines discriminative object regions with classification networks to address weakly‑supervised semantic segmentation. The method starts from a single small region, iteratively erases mined areas to force the classifier to uncover new complementary regions, and augments this process with an online prohibitive segmentation learner that supplies auxiliary supervision guided by reliable classification scores, yielding a dense, complete object mask for segmentation. The approach attains 55.0% and 55.7% mIoU on PASCAL VOC 2012 val and test sets, surpassing previous state‑of‑the‑art results.

Abstract

We investigate a principle way to progressively mine discriminative object regions using classification networks to address the weakly-supervised semantic segmentation problems. Classification networks are only responsive to small and sparse discriminative regions from the object of interest, which deviates from the requirement of the segmentation task that needs to localize dense, interior and integral regions for pixel-wise inference. To mitigate this gap, we propose a new adversarial erasing approach for localizing and expanding object regions progressively. Starting with a single small object region, our proposed approach drives the classification network to sequentially discover new and complement object regions by erasing the current mined regions in an adversarial manner. These localized regions eventually constitute a dense and complete object region for learning semantic segmentation. To further enhance the quality of the discovered regions by adversarial erasing, an online prohibitive segmentation learning approach is developed to collaborate with adversarial erasing by providing auxiliary segmentation supervision modulated by the more reliable classification scores. Despite its apparent simplicity, the proposed approach achieves 55.0% and 55.7% mean Intersection-over-Union (mIoU) scores on PASCAL VOC 2012 val and test sets, which are the new state-of-the-arts.