On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption
This work addresses robustness for safety-critical applications like semantic segmentation, but it is incremental as it builds on existing IBE techniques.
The paper tackled the problem of improving robustness to input corruptions in semantic segmentation models by analyzing optimization objectives like Softmax, IBE, and Sigmoid. As a result, they proposed SCrIBE, which combines Sigmoid with IBE, achieving a mIOU of 42.1 compared to 40.3 for IBE and 37.5 for the baseline.
Semantic segmentation is a scene understanding task at the heart of safety-critical applications where robustness to corrupted inputs is essential. Implicit Background Estimation (IBE) has demonstrated to be a promising technique to improve the robustness to out-of-distribution inputs for semantic segmentation models for little to no cost. In this paper, we provide analysis comparing the structures learned as a result of optimization objectives that use Softmax, IBE, and Sigmoid in order to improve understanding their relationship to robustness. As a result of this analysis, we propose combining Sigmoid with IBE (SCrIBE) to improve robustness. Finally, we demonstrate that SCrIBE exhibits superior segmentation performance aggregated across all corruptions and severity levels with a mIOU of 42.1 compared to both IBE 40.3 and the Softmax Baseline 37.5.