FoodSeg-103
Emerging2papers using it
2025first seen
FoodSeg-103 is a dataset containing 2,135 food images used to evaluate ingredient-level semantic segmentation models by predicting an ingredient class for each pixel in the images.
FoodSeg-103 is a dataset containing 2,135 food images used to evaluate ingredient-level semantic segmentation models by predicting an ingredient class for each pixel in the images.