# Jumpat/SegmentAnythingin3D

Segment Anything in 3D with NeRFs (NeurIPS 2023 & IJCV 2025)

Repository: https://github.com/Jumpat/SegmentAnythingin3D
Canonical: https://ross.abutalabs.com/products/segmentanythingin3d
Language: Python
License: Apache-2.0
License Family: permissive
Topics: 3d, 3d-segmentation, computer-vision, nerf, segment-anything, segmentation, deep-learning
Last push: 2025-05-19T02:37:14+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 22, release rhythm 35, longevity 87
- inputs: {"age_days": 1227, "days_push": 472, "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 1030, forks 64 (observed 2026-08-28T04:03:17.861224+00:00)

## What it is
SA3D is a research framework that lifts 2D Segment Anything (SAM) masks into 3D segmentation of objects within a NeRF or 3D Gaussian Splatting scene from a single prompted view. It includes a NeRF-based implementation, a nerfstudio variant, and a 3D-GS segmentation dataset.

## Use cases
- segment an object in 3d from a nerf scene
- get 3d masks from segment anything prompts
- extract a 3d model of a target object from a radiance field
- segment anything in 3d with gaussian splatting
- research on 3d scene perception and segmentation
- one-shot 3d segmentation from a single view

## When to choose
- you already have a NeRF or 3D-GS representation of a scene and need object-level 3D segmentation
- you want a published, peer-reviewed method (NeurIPS 2023 / IJCV 2025) for SAM-based 3D lifting
- you need a benchmark dataset for 3D Gaussian Splatting segmentation

## When to avoid
- you need segmentation of raw images or videos without an existing NeRF/3D-GS model
- you need a production-ready, optimized pipeline rather than research code
- your project does not involve neural radiance fields or 3D scenes

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, image-processing, machine-learning, deep-learning
- domain: computer-vision, deep-learning, machine-learning, artificial-intelligence
- platform: python
- tags: 3d-segmentation, nerf, segment-anything, gaussian-splatting, research, neurips, linux, gpu

## Member repositories
- Jumpat/SegmentAnythingin3D (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:17.861224+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 2026-08-30T07:07:31.221763+00:00, confidence not recorded.
  - readme: https://github.com/Jumpat/SegmentAnythingin3D (fetched 2026-08-28T04:03:17.861224+00:00, sha 8448b79bea92)
- Data as of 2026-08-30T08:39:29.467469+00:00.