# isl-org/MiDaS

Code for robust monocular depth estimation described in "Ranftl et. al., Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer, TPAMI 2022"

Repository: https://github.com/isl-org/MiDaS
Canonical: https://ross.abutalabs.com/products/midas
Language: Python
License: MIT
License Family: permissive
Topics: monocular-depth-estimation, single-image-depth-prediction, deeplearning
Archived: true
Last push: 2024-08-23T10:21:06+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2627, "days_push": 740, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5420, forks 724 (observed 2026-08-28T04:09:17.909552+00:00)

## What it is
MiDaS is a Python library with pretrained models for robust monocular depth estimation from a single image, based on the TPAMI 2022 paper and Vision Transformers for Dense Prediction. It offers multiple model variants trading off quality and speed, including small models for embedded devices and OpenVINO support for Intel CPUs.

## Use cases
- estimate depth from a single photo
- run monocular depth estimation on embedded devices
- add depth maps to a computer vision pipeline
- compare depth estimation model quality vs speed
- use pretrained depth models without training data

## When to choose
- you need state-of-the-art zero-shot depth estimation from single images
- you want a range of model sizes from embedded to high accuracy
- you want MIT-licensed pretrained depth models

## When to avoid
- you need metric depth in real-world units rather than relative depth
- you need stereo or multi-view depth from multiple cameras
- you need real-time depth on CPU without OpenVINO-compatible hardware

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, machine-learning, deep-learning, robotics
- platform: python, cross-platform
- tags: monocular-depth-estimation, depth-prediction, vision-transformer, zero-shot-transfer, pretrained-models, pytorch, gpu

## Member repositories
- isl-org/MiDaS (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:17.909552+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 2026-08-29T17:56:56.753996+00:00, confidence not recorded.
  - readme: https://github.com/isl-org/MiDaS (fetched 2026-08-28T04:09:17.909552+00:00, sha f339d58e6a25)
- Data as of 2026-08-30T08:39:29.467469+00:00.