Mars-Bench: mb-crater_multi_seg

The mb-crater_multi_seg dataset provides multi-class crater segmentation using THEMIS imagery, where each pixel is assigned to one of four crater types: Other, Layered, Buried, or Secondary. This dataset supports detailed morphological classification and enables analysis of how distinct crater types contribute to understanding Martian surface evolution. Craters are crucial for understanding the geological history, surface age, and impact processes of planetary bodies. Moreover, classifying craters into distinct morphological types allows researchers to assess which crater categories are most informative for scientific interpretation. Although craters are often described as bowl-shaped, they are not always perfectly circular. To address this, we provide annotations in elliptical form, marking the first known release of crater segmentation using elliptical geometry, rather than the circular annotations common in prior datasets. As with the binary dataset, the full image-mask pairs were generated from open-source THEMIS data, restricted to approximately ±30° latitude to mitigate missing polar pixels. Geographically disjoint splits ensure no spatial overlap: −180° – 60° for training, 60° – 120° for testing, and 120° – 180° for validation. To explore all Mars-Bench resources, visit our project website: https://mars-bench.github.io/