# charlesq34/frustum-pointnets

Frustum PointNets for 3D Object Detection from RGB-D Data

Repository: https://github.com/charlesq34/frustum-pointnets
Canonical: https://ross.abutalabs.com/products/frustum-pointnets
Language: Python
License: Apache-2.0
License Family: permissive
Topics: object-detection, 3d, point-cloud, robotics, deep-learning
Last push: 2020-03-24T20:42:38+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3072, "days_push": 2353, "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 1668, forks 530 (observed 2026-08-28T04:05:19.816331+00:00)

## What it is
Official TensorFlow code release for the CVPR 2018 paper 'Frustum PointNets for 3D Object Detection from RGB-D Data' by Stanford and Nuro researchers. It combines 2D object detectors with PointNet/PointNet++ networks to perform 3D instance segmentation and amodal 3D bounding box estimation directly on point clouds.

## Use cases
- detect 3d objects from rgb-d data
- run 3d object detection on kitti benchmark
- train frustum pointnet on sunrgbd dataset
- segment 3d instances in point clouds
- estimate amodal 3d bounding boxes from lidar point clouds
- reproduce cvpr 2018 3d detection research results

## When to choose
- you need a proven 3D object detection pipeline for RGB-D or LiDAR data
- you want to reproduce or build on the Frustum PointNets paper
- you work with KITTI or SUN RGB-D datasets and TensorFlow
- you need direct point-cloud processing without voxelization

## When to avoid
- you need a maintained production system with recent framework support
- you prefer PyTorch over TensorFlow 1.x
- you need state-of-the-art 3D detection beyond 2018-era methods
- you want plug-and-play inference without research setup effort

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, robotics, autonomous-vehicles, deep-learning, artificial-intelligence
- platform: python
- tags: 3d-object-detection, point-cloud, pointnet, rgb-d, kitti, sunrgbd, cvpr-2018, tensorflow, instance-segmentation, research-code, linux, gpu

## Member repositories
- charlesq34/frustum-pointnets (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:19.816331+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 2026-08-30T03:42:49.427828+00:00, confidence not recorded.
  - readme: https://github.com/charlesq34/frustum-pointnets (fetched 2026-08-28T04:05:19.816331+00:00, sha b7e2addbdd25)
- Data as of 2026-08-30T08:39:29.467469+00:00.