Oneformer Coco Swin Large
OneFormer is the first multi-task universal image segmentation framework, achieving semantic segmentation, instance segmentation, and panoptic segmentation tasks with a single model
Downloads 165.70k
Release Time : 11/15/2022
Model Overview
This model is a large OneFormer model trained on the COCO dataset, utilizing the Swin backbone network. As a unified image segmentation framework, it surpasses specialized models across multiple segmentation tasks with just one model.
Model Features
Unified Multi-task Architecture
The first framework to achieve semantic/instance/panoptic segmentation with a single model, eliminating the need for specialized models
Dynamic Task Inference
Guides the model to focus on the current task through task tokens, enabling task-oriented training and dynamic task inference
Surpasses Specialized Models
On the COCO dataset, the single model outperforms traditional specialized models in all segmentation tasks
Model Capabilities
Semantic Segmentation
Instance Segmentation
Panoptic Segmentation
Image Scene Understanding
Object Recognition and Localization
Use Cases
Computer Vision
Autonomous Driving Scene Parsing
Used for semantic segmentation of road scenes, identifying elements such as roads, pedestrians, and vehicles
Generates precise scene segmentation maps
Medical Image Analysis
Performs instance segmentation of organs or lesion areas in medical images
Assists doctors in quantitative analysis
Remote Sensing Image Processing
Conducts panoptic segmentation of satellite/aerial images to identify various land cover types
Supports land cover classification and change detection
Featured Recommended AI Models
Š 2025AIbase