SlimSeg: Slimmable Semantic Segmentation with Boundary Supervision
This addresses the need for flexible, efficient semantic segmentation in practical applications, representing an incremental improvement over existing lightweight models.
The paper tackles the problem of inflexible computational cost in semantic segmentation models by proposing SlimSeg, a slimmable method that adapts to varying accuracy-efficiency tradeoffs, achieving better performance than independent models on benchmarks like Cityscapes and CamVid.
Accurate semantic segmentation models typically require significant computational resources, inhibiting their use in practical applications. Recent works rely on well-crafted lightweight models to achieve fast inference. However, these models cannot flexibly adapt to varying accuracy and efficiency requirements. In this paper, we propose a simple but effective slimmable semantic segmentation (SlimSeg) method, which can be executed at different capacities during inference depending on the desired accuracy-efficiency tradeoff. More specifically, we employ parametrized channel slimming by stepwise downward knowledge distillation during training. Motivated by the observation that the differences between segmentation results of each submodel are mainly near the semantic borders, we introduce an additional boundary guided semantic segmentation loss to further improve the performance of each submodel. We show that our proposed SlimSeg with various mainstream networks can produce flexible models that provide dynamic adjustment of computational cost and better performance than independent models. Extensive experiments on semantic segmentation benchmarks, Cityscapes and CamVid, demonstrate the generalization ability of our framework.