facebookresearch/meshrcnn

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1161 stars · 175 forks observed · 2026-08-28

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

Mesh R-CNN is Facebook AI Research's official implementation of the ICCV 2019 paper, a model that detects objects in images and predicts their 3D shape as triangle meshes. It is built on top of Detectron2 and PyTorch3D and includes demos, training, and evaluation code for the Pix3D and ShapeNet benchmarks.

Use cases

• predict 3D mesh models of objects from a single image
• run 3D object detection and instance segmentation on photos
• reproduce Mesh R-CNN results on the Pix3D benchmark
• train and evaluate 3D shape prediction on ShapeNet
• experiment with mesh prediction heads in a Detectron2 pipeline

When to choose

• you need research-grade code for image-to-3D mesh prediction
• your project already uses Detectron2 and PyTorch3D
• you want to reproduce or extend the Mesh R-CNN paper

When to avoid

• you need a production-ready, actively supported 3D vision product
• you want a simple pretrained API without installing Detectron2 and PyTorch3D
• you need real-time 3D reconstruction on edge devices

Facets

library · maturity maintenance

machine-learning deep-learning computer-vision image-processing computer-vision machine-learning deep-learning python 3d-reconstruction mesh-prediction instance-segmentation detectron2 pytorch3d iccv-2019 research-code research linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="facebookresearch/meshrcnn")