Distribution-aware Margin Calibration for Semantic Segmentation in Images
This work addresses a key challenge in semantic segmentation for computer vision researchers, offering a theoretically grounded method to enhance IoU performance, though it is incremental as it builds on existing surrogate optimization approaches.
The paper tackles the problem of optimizing the Jaccard index (IoU) for semantic segmentation by proposing a margin calibration method that provides a theoretical lower bound for improved generalization, resulting in substantial IoU improvements on seven image datasets.
The Jaccard index, also known as Intersection-over-Union (IoU), is one of the most critical evaluation metrics in image semantic segmentation. However, direct optimization of IoU score is very difficult because the learning objective is neither differentiable nor decomposable. Although some algorithms have been proposed to optimize its surrogates, there is no guarantee provided for the generalization ability. In this paper, we propose a margin calibration method, which can be directly used as a learning objective, for an improved generalization of IoU over the data-distribution, underpinned by a rigid lower bound. This scheme theoretically ensures a better segmentation performance in terms of IoU score. We evaluated the effectiveness of the proposed margin calibration method on seven image datasets, showing substantial improvements in IoU score over other learning objectives using deep segmentation models.