MaskDINOModelSize
Enums · AiDotNet.Enums
Defines the backbone size variants for Mask DINO.
For Beginners
Mask DINO unifies object detection and segmentation by extending the DINO detector with a mask prediction branch. It handles instance, panoptic, and semantic segmentation in one architecture.
How It Works
Technical Details: Built on DINO (DETR with Improved deNoising anchOr boxes) with an additional mask branch. Uses ResNet or Swin Transformer backbones with deformable attention in the transformer encoder-decoder.
Reference: Li et al., “Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation”, CVPR 2023.
Fields
| Field | Summary |
|---|---|
MaskDINOModelSize R50 |
ResNet-50 backbone (44M params). |
MaskDINOModelSize SwinLarge |
Swin-L backbone (218M params). |
int value__ |