RoadFormer+: Delivering RGB-X Scene Parsing through Scale-Aware Information Decoupling and Advanced Heterogeneous Feature Fusion

Benchmark Model Rank Results
semantic-segmentation-on-fmb-datasetRoadFormer+ (RGB-Infrared)mIoU: 73.1
semantic-segmentation-on-syn-udtiriRoadFormer+ (B)IoU: 94.11
semantic-segmentation-on-zju-rgb-pRoadFormer+ (ConvNeXt-B, RGB-AoLP)mIoU: 92.9
semantic-segmentation-on-zju-rgb-pRoadFormer+ (ConvNeXt-L, RGB-AoLP)mIoU: 93.0
thermal-image-segmentation-on-mfn-datasetRoadFormer+ (ConvNeXt-L)mIOU: 62.7