Semantic Segmentation with Labeling Uncertainty and Class Imbalance
This paper tackles the problem of improving robustness and accuracy in semantic segmentation for practitioners dealing with noisy or imbalanced datasets, offering an incremental improvement to existing methods.
This paper addresses semantic segmentation challenges, specifically class imbalance and pixel-labeling uncertainty. The authors propose a method that assigns a weight to each pixel based on its class and labeling uncertainty, which is then used during training to adjust pixel importance. This approach significantly improves performance in three segmentation tasks and demonstrates increased noise invariance.
Recently, methods based on Convolutional Neural Networks (CNN) achieved impressive success in semantic segmentation tasks. However, challenges such as the class imbalance and the uncertainty in the pixel-labeling process are not completely addressed. As such, we present a new approach that calculates a weight for each pixel considering its class and uncertainty during the labeling process. The pixel-wise weights are used during training to increase or decrease the importance of the pixels. Experimental results show that the proposed approach leads to significant improvements in three challenging segmentation tasks in comparison to baseline methods. It was also proved to be more invariant to noise. The approach presented here may be used within a wide range of semantic segmentation methods to improve their robustness.