# tg-bomze/Face-Depixelizer

Face Depixelizer based on "PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models" repository.

Repository: https://github.com/tg-bomze/Face-Depixelizer
Canonical: https://ross.abutalabs.com/products/face-depixelizer
Language: Jupyter Notebook
License Family: other
Last push: 2025-01-06T11:27:36+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": 2266, "days_push": 604, "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 2560, forks 294 (observed 2026-08-28T04:07:00.907503+00:00)

## What it is
A Jupyter Notebook-based tool that turns low-resolution pixelated face images into realistic high-resolution faces using the PULSE method, which searches StyleGAN's latent space for images that downscale to the input. It is designed to run on Google Colab and produces imaginary faces, not reconstructions of real people.

## Use cases
- upscale a pixelated face photo to high resolution
- depixelize blurred face images
- generate realistic faces from low-resolution inputs
- explore StyleGAN latent space for face synthesis
- run PULSE face upsampling in Google Colab

## When to choose
- you have a very low-resolution or pixelated face image and want a photorealistic high-resolution version
- you want a ready-to-run Colab notebook without local GPU setup
- you are experimenting with GAN-based face upsampling

## When to avoid
- you need to identify or reconstruct the actual original person - the output is an imaginary face, not the real one
- you need a production-ready tool with a license or stable API
- you cannot tolerate Google Drive download quota limits for model weights

## Facets
- artifact type: application
- maturity: maintenance
- function: image-processing, machine-learning, deep-learning
- domain: computer-vision, image-processing, deep-learning, artificial-intelligence
- platform: python, cross-platform
- tags: face-depixelizer, super-resolution, stylegan, pulse, google-colab, jupyter-notebook, face-upscaling, generative-models, gpu

## Member repositories
- tg-bomze/Face-Depixelizer (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:00.907503+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 2026-08-30T02:23:31.764243+00:00, confidence not recorded.
  - readme: https://github.com/tg-bomze/Face-Depixelizer (fetched 2026-08-28T04:07:00.907503+00:00, sha 92658f8197d6)
- Data as of 2026-08-30T08:39:29.467469+00:00.