Uncertainty estimates for semantic segmentation: providing enhanced reliability for automated motor claims handling
This work addresses reliability issues for insurance companies using automated damage assessment, but it is incremental as it applies existing uncertainty estimation techniques to a specific domain.
The paper tackled the problem of unreliable semantic segmentation outputs in automated motor claims handling by using a meta-classification model to filter out low-quality segments, resulting in a 16 percentage point improvement in mIoU and a 77% reduction in wrongly predicted segments.
Deep neural network models for image segmentation can be a powerful tool for the automation of motor claims handling processes in the insurance industry. A crucial aspect is the reliability of the model outputs when facing adverse conditions, such as low quality photos taken by claimants to document damages. We explore the use of a meta-classification model to empirically assess the precision of segments predicted by a model trained for the semantic segmentation of car body parts. Different sets of features correlated with the quality of a segment are compared, and an AUROC score of 0.915 is achieved for distinguishing between high- and low-quality segments. By removing low-quality segments, the average mIoU of the segmentation output is improved by 16 percentage points and the number of wrongly predicted segments is reduced by 77%.