Segment Every Out-of-Distribution Object
About
Semantic segmentation models, while effective for in-distribution categories, face challenges in real-world deployment due to encountering out-of-distribution (OoD) objects. Detecting these OoD objects is crucial for safety-critical applications. Existing methods rely on anomaly scores, but choosing a suitable threshold for generating masks presents difficulties and can lead to fragmentation and inaccuracy. This paper introduces a method to convert anomaly \textbf{S}core \textbf{T}o segmentation \textbf{M}ask, called S2M, a simple and effective framework for OoD detection in semantic segmentation. Unlike assigning anomaly scores to pixels, S2M directly segments the entire OoD object. By transforming anomaly scores into prompts for a promptable segmentation model, S2M eliminates the need for threshold selection. Extensive experiments demonstrate that S2M outperforms the state-of-the-art by approximately 20% in IoU and 40% in mean F1 score, on average, across various benchmarks including Fishyscapes, Segment-Me-If-You-Can, and RoadAnomaly datasets.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Anomaly Segmentation | Fishyscapes Lost & Found (val) | -- | 53 | |
| Out-of-Distribution Detection | SMIYC Anomaly (val) | AuPRC91.92 | 9 | |
| Out-of-Distribution Detection | SMIYC Obstacle (val) | AuPRC91.73 | 9 | |
| Anomaly Segmentation | SMIYC Anomaly (val) | AuIoU54.33 | 5 | |
| Anomaly Segmentation | SMIYC Obstacle (val) | AuIoU58.12 | 5 | |
| Anomaly Segmentation | RoadAnomaly (val) | AuIoU54.31 | 5 | |
| Open-Set Semantic Segmentation | FS Static | IoU70 | 5 | |
| Open-Set Semantic Segmentation | Lost & Found FS | IoU30.5 | 5 | |
| Open-Set Semantic Segmentation | SMIYC Anomaly | IoU77.5 | 5 | |
| Open-Set Semantic Segmentation | SMIYC-Obstacle | IoU67.6 | 5 |