YCOR Dataset
A semantic segmentation dataset for off-road scenes, constructed through a collaboration between Yamaha and Carnegie Mellon University. It is designed for unstructured outdoor environments and includes categories such as grass, dirt, shrubs, rocks, and water.
Timeline (last 90 days)
After training on the Yamaha-CMU Off-Road (YCOR) dataset, SegFormer-B2 achieves a validation set mIoU of approximately 57.4%.
The YCOR dataset is designed specifically for unstructured off-road scenes, containing off-road categories such as grass, dirt, shrubs, rocks, and water, and is one of the most commonly used public benchmarks in the field of off-road semantic segmentation.
All Facts (2)
After training on the Yamaha-CMU Off-Road (YCOR) dataset, SegFormer-B2 achieves a validation set mIoU of approximately 57.4%.
50%UnverifiedThe YCOR dataset is designed specifically for unstructured off-road scenes, containing off-road categories such as grass, dirt, shrubs, rocks, and water, and is one of the most commonly used public benchmarks in the field of off-road semantic segmentation.
50%