# chrischoy/3D-R2N2

Single/multi view image(s) to voxel reconstruction using a recurrent neural network

Repository: https://github.com/chrischoy/3D-R2N2
Canonical: https://ross.abutalabs.com/products/3d-r2n2
Language: Python
License: MIT
License Family: permissive
Last push: 2021-07-16T02:33:14+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3679, "days_push": 1875, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1410, forks 291 (observed 2026-08-28T04:04:38.760404+00:00)

## What it is
3D-R2N2 is a PyTorch-based implementation of a recurrent neural network that reconstructs voxelized 3D models of objects from one or multiple 2D images. It accompanies the ECCV 2016 paper and introduced the 3D Convolutional LSTM for order-invariant multi-view reconstruction.

## Use cases
- reconstruct a 3D voxel model from a single photo
- generate 3D object reconstructions from multiple unordered views
- reproduce the 3D-R2N2 ECCV 2016 paper results
- train a 3D convolutional LSTM on ShapeNet data
- convert e-commerce product images into voxelized 3D shapes

## When to choose
- you need a research baseline for learning-based single/multi-view 3D reconstruction
- you want voxelized outputs from arbitrary numbers of input images
- you are studying 3D convolutional LSTM/GRU architectures

## When to avoid
- you need high-resolution meshes rather than coarse voxel grids
- you want actively maintained code with modern framework support
- you need production-ready 3D reconstruction rather than a research prototype

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning
- platform: python
- tags: 3d-reconstruction, voxel, rnn, lstm, single-view-reconstruction, multi-view-reconstruction, shapenet, eccv-2016, research, linux, gpu

## Member repositories
- chrischoy/3D-R2N2 (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:38.760404+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:38:28.116162+00:00, confidence not recorded.
  - readme: https://github.com/chrischoy/3D-R2N2 (fetched 2026-08-28T04:04:38.760404+00:00, sha 2464d8116ba0)
- Data as of 2026-08-30T08:39:29.467469+00:00.