Small Objects Matters in Weakly-supervised Semantic Segmentation
This work addresses a specific limitation in weakly-supervised semantic segmentation for computer vision researchers, but it is incremental as it builds on existing methods with a targeted improvement.
The paper tackled the problem of weakly-supervised semantic segmentation by revealing that existing methods struggle with small objects, and proposed a size-balanced cross-entropy loss that improved performance across ten baselines on three datasets.
Weakly-supervised semantic segmentation (WSSS) performs pixel-wise classification given only image-level labels for training. Despite the difficulty of this task, the research community has achieved promising results over the last five years. Still, current WSSS literature misses the detailed sense of how well the methods perform on different sizes of objects. Thus we propose a novel evaluation metric to provide a comprehensive assessment across different object sizes and collect a size-balanced evaluation set to complement PASCAL VOC. With these two gadgets, we reveal that the existing WSSS methods struggle in capturing small objects. Furthermore, we propose a size-balanced cross-entropy loss coupled with a proper training strategy. It generally improves existing WSSS methods as validated upon ten baselines on three different datasets.