Improving Weakly-supervised Object Localization via Causal Intervention
This addresses a key limitation in weakly supervised object localization for computer vision applications, though it is an incremental improvement over existing methods.
The paper tackles the problem of weakly supervised object localization being confounded by co-occurring contexts (e.g., bird and sky) by using causal intervention to eliminate this bias, resulting in significant accuracy improvements, such as 58.39% vs. 52.4% top-1 localization accuracy on CUB-200-2011.
The recent emerged weakly supervised object localization (WSOL) methods can learn to localize an object in the image only using image-level labels. Previous works endeavor to perceive the interval objects from the small and sparse discriminative attention map, yet ignoring the co-occurrence confounder (e.g., bird and sky), which makes the model inspection (e.g., CAM) hard to distinguish between the object and context. In this paper, we make an early attempt to tackle this challenge via causal intervention (CI). Our proposed method, dubbed CI-CAM, explores the causalities among images, contexts, and categories to eliminate the biased co-occurrence in the class activation maps thus improving the accuracy of object localization. Extensive experiments on several benchmarks demonstrate the effectiveness of CI-CAM in learning the clear object boundaries from confounding contexts. Particularly, in CUB-200-2011 which severely suffers from the co-occurrence confounder, CI-CAM significantly outperforms the traditional CAM-based baseline (58.39% vs 52.4% in top-1 localization accuracy). While in more general scenarios such as ImageNet, CI-CAM can also perform on par with the state of the arts.