Few Shot Image Segmentation
Few-shot semantic segmentation (FSS) is a subtask in the field of computer vision that aims to learn to segment target objects in query images using only a few pixel-level annotated support images. This task enhances the model's ability to quickly adapt to new categories by reducing the need for labeled data, making it particularly valuable in applications such as medical image analysis, autonomous driving, and remote sensing image processing.
COCO-20i (1-shot)
COCO-20i (10-shot)
DGPNet (ResNet-101)
COCO-20i (2-way 1-shot)
Label Anything (Vit-B/16-SAM)
COCO-20i (5-shot)
COCO-20i -> Pascal VOC (1-shot)
MSDNet (ResNet-101)
COCO-20i -> Pascal VOC (5-shot)
FPTrans (DeiT-B/16)
FSS-1000
LSeg
FSS-1000 (1-shot)
VAT
FSS-1000 (5-shot)
EfficientLab + PRN
PASCAL-5i (1-Shot)
SegGPT (ViT)
PASCAL-5i (10-Shot)
PASCAL-5i (5-Shot)
SegGPT (ViT)
Pascal5i