Vision Models
Notable detection, segmentation and foundation models.
This is an overview — pick a sub-topic below.
Sub-topics (5)
Segment Anything (SAM, SAM 2, FastSAM, MobileSAM)
Promptable segmentation models that cut out any object from a click, box or text (with Grounding DINO).
Grounding DINO, Florence-2 and DINOv2
Open-vocabulary detection from text prompts (Grounding DINO), a multi-task vision model (Florence-2), and strong self-supervised features (DINOv2).
DETR and RT-DETR
Transformer-based detectors; RT-DETR is a real-time variant competitive with YOLO.
R-CNN Family, RetinaNet, SSD, EfficientDet and CenterNet
Classic detector architectures: two-stage (Faster/Mask R-CNN) and one-stage (RetinaNet, SSD, EfficientDet, CenterNet).
U-Net, DeepLab and SegFormer
Segmentation architectures: U-Net (encoder-decoder with skip connections), DeepLab (atrous convolution), SegFormer (Transformer).
Spotted a mistake or want to add detail? Suggest an improvement — Content Contributors' edits are reviewed by editors before they go live.
Discussion (0)
Sign in or join free to ask questions and discuss this topic.