msracver/FCIS

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1561 stars · 403 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

FCIS is the official MXNet implementation of the CVPR 2017 paper 'Fully Convolutional Instance-aware Semantic Segmentation', which won first place in the COCO 2016 segmentation challenge. It provides a fast, end-to-end fully convolutional framework that jointly performs instance mask estimation and categorization.

Use cases

• segment individual object instances in images
• reproduce COCO 2016 challenge-winning instance segmentation results
• compare instance segmentation baselines against Mask R-CNN
• research fully convolutional approaches to instance-aware segmentation
• train and evaluate instance segmentation models on COCO with MXNet

When to choose

• you need a fast, fully convolutional instance segmentation baseline
• you are reproducing or building on the FCIS CVPR 2017 paper
• your stack is already MXNet-based and you need instance segmentation

When to avoid

• you need actively maintained code or modern framework support (PyTorch/TensorFlow)
• you want state-of-the-art accuracy with modern tricks like FPN and ROIAlign (Mask R-CNN style)
• you need production-ready segmentation with long-term support

Facets

library · maturity maintenance

machine-learning deep-learning computer-vision image-processing computer-vision deep-learning machine-learning image-processing python instance-segmentation semantic-segmentation mxnet coco cvpr-2017 research-code microsoft-research linux gpu cuda

Member repositories

For agents

markdown · JSON · MCP: product_card(name="msracver/FCIS")