# torrvision/crfasrnn

This repository contains the source code for the semantic image segmentation method described in the ICCV 2015 paper: Conditional Random Fields as Recurrent Neural Networks. http://crfasrnn.torr.vision/

Repository: https://github.com/torrvision/crfasrnn
Canonical: https://ross.abutalabs.com/products/crfasrnn
Language: MATLAB
License: NOASSERTION
License Family: other
Last push: 2019-11-30T13:17:39+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3990, "days_push": 2468, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1335, forks 455 (observed 2026-08-28T04:04:25.337618+00:00)

## What it is
Reference implementation of CRF-RNN, an ICCV 2015 semantic image segmentation method that integrates conditional random fields into a neural network, built on the Caffe deep learning library. It includes a trained model recognizing 20 object classes and links to PyTorch and Keras ports.

## Use cases
- semantic image segmentation of objects in photos
- run CRF-RNN segmentation model on my own images
- integrate CRF layers into a CNN for pixel-wise labeling
- reproduce ICCV 2015 CRF as RNN paper results
- segment 20 PASCAL VOC object classes in images

## When to choose
- you need the original CRF-RNN implementation or its trained Caffe model
- you are reproducing or extending the ICCV 2015 paper
- you want to study how CRFs are embedded in deep networks

## When to avoid
- you want a modern maintained framework - use the PyTorch or Keras ports instead
- you need segmentation of classes beyond the 20 trained ones without retraining
- your stack does not support the legacy Caffe/MATLAB toolchain

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, image-processing, machine-learning
- platform: cpp, python
- tags: semantic-segmentation, conditional-random-fields, caffe, crf-rnn, research-code, matlab, linux, gpu

## Member repositories
- torrvision/crfasrnn (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:25.337618+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 2026-08-30T04:44:25.445138+00:00, confidence not recorded.
  - readme: https://github.com/torrvision/crfasrnn (fetched 2026-08-28T04:04:25.337618+00:00, sha 7f2f3fb43e1c)
- Data as of 2026-08-30T08:39:29.467469+00:00.