ICRA

GOOSE-M2F: Adapting Mask2Former for High-Fidelity, Long-Tailed Fine-Grained Semantic Segmentation in Unstructured Outdoor Terrain

GOOSE-M2F：适配Mask2Former用于非结构化户外地形的高保真、长尾细粒度语义分割

Jyothiraditya Lingam, Nikhileswara Rao Sulake, Sai Manikanta Eswar Machara

机构 * Rajiv Gandhi University of Knowledge Technologies, Nuzvid, India（拉吉夫·甘地知识技术大学，努兹维德，印度） 纠错

AI总结 针对非结构化户外地形长尾细粒度语义分割挑战，提出GOOSE-M2F，通过200个对象查询、特征精炼模块和辅助监督头，结合多阶段训练策略，在GOOSE基准上达到70.08%复合mIoU。

Comments This solution has got 3rd position at GOOSE 2D Fine-Grained Semantic Segmentation (FGSS) Challenge at ICRA~2026