# luanfujun/deep-painterly-harmonization

Code and data for paper "Deep Painterly Harmonization": https://arxiv.org/abs/1804.03189

Repository: https://github.com/luanfujun/deep-painterly-harmonization
Canonical: https://ross.abutalabs.com/products/deep-painterly-harmonization
Language: Cuda
License Family: other
Last push: 2021-08-02T02:46:03+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": 3068, "days_push": 1857, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6042, forks 607 (observed 2026-08-28T04:09:34.549493+00:00)

## What it is
Research code implementing the 'Deep Painterly Harmonization' algorithm, which seamlessly blends a pasted object into a painting's style using deep neural features. It is built on Torch7 with CUDA and includes data and scripts to reproduce the paper's results.

## Use cases
- harmonize a pasted object into a painting's style
- reproduce results from the Deep Painterly Harmonization paper
- blend an inserted element so it matches a painting's brushwork
- experiment with neural image harmonization techniques
- generate stylized composites from paintings and photos

## When to choose
- you need to composite an object into a painting so it looks naturally painted
- you want to reproduce or build on the paper's research
- you have a CUDA-capable Linux setup and are comfortable with Torch7 and Matlab/Octave

## When to avoid
- you need a maintained, production-ready tool - the code is old research software with no license
- you want a modern PyTorch or GPU-friendly pipeline
- you need commercial use - the software is restricted to academic and non-commercial use
- you want automatic weight estimation - the pretrained stylization model was never released

## Facets
- artifact type: library
- maturity: maintenance
- function: image-processing, deep-learning, machine-learning
- domain: computer-vision, image-processing, deep-learning, artificial-intelligence
- platform: python, cpp
- tags: neural-style-transfer, image-harmonization, torch7, vgg-19, research-code, cuda, academic-paper, non-commercial, linux, gpu

## Member repositories
- luanfujun/deep-painterly-harmonization (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:34.549493+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 2026-08-29T17:48:57.667510+00:00, confidence not recorded.
  - readme: https://github.com/luanfujun/deep-painterly-harmonization (fetched 2026-08-28T04:09:34.549493+00:00, sha 9c55a2fa3b38)
- Data as of 2026-08-30T08:39:29.467469+00:00.