# facebookresearch/consistent_depth

We estimate dense, flicker-free, geometrically consistent depth from monocular video, for example hand-held cell phone video.

Repository: https://github.com/facebookresearch/consistent_depth
Canonical: https://ross.abutalabs.com/products/consistent_depth
Language: Python
License: MIT
License Family: permissive
Archived: true
Last push: 2023-07-23T15:08:54+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2311, "days_push": 1137, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1634, forks 227 (observed 2026-08-28T04:05:14.312954+00:00)

## What it is
A research library from Facebook AI Research implementing Consistent Video Depth Estimation (SIGGRAPH 2020). It reconstructs dense, flicker-free, geometrically consistent depth maps for all pixels in monocular video, such as hand-held cell phone footage, by fine-tuning a single-image depth network against structure-from-motion constraints.

## Use cases
- estimate depth maps from monocular video
- make flicker-free depth for handheld phone video
- reconstruct scenes from a single camera video
- create video visual effects using depth
- convert 2D video to 3D-aware content
- run depth estimation on a Colab notebook

## When to choose
- you need temporally consistent depth for a specific monocular clip and can afford per-video optimization
- you have an NVIDIA GPU and can install COLMAP and FFmpeg
- you want a research-grade reference implementation of the SIGGRAPH 2020 paper

## When to avoid
- you need real-time depth estimation
- you want a plug-and-play pretrained model without per-video fine-tuning
- you have no GPU or cannot set up COLMAP dependencies
- you need actively maintained production software

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing, video-processing
- domain: computer-vision, machine-learning, deep-learning
- platform: python
- tags: depth-estimation, monocular-video, structure-from-motion, siggraph-2020, research-code, 3d-reconstruction, video, linux, gpu

## Member repositories
- facebookresearch/consistent_depth (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:14.312954+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 2026-08-30T03:47:03.428731+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/consistent_depth (fetched 2026-08-28T04:05:14.312954+00:00, sha d7a68b3cbe83)
- Data as of 2026-08-30T08:39:29.467469+00:00.