# atriumlts/subpixel

subpixel: A subpixel convnet for super resolution with Tensorflow

Repository: https://github.com/atriumlts/subpixel
Canonical: https://ross.abutalabs.com/products/subpixel
Language: Jupyter Notebook
License: MIT
License Family: permissive
Last push: 2022-05-18T18:27:49+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": 3674, "days_push": 1568, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2123, forks 299 (observed 2026-08-28T04:06:16.700548+00:00)

## What it is
A TensorFlow reimplementation of the efficient sub-pixel convolutional neural network (ESPCN) for single-image super-resolution, based on Shi et al.'s CVPR 2016 paper. It includes training notebooks and example 4x upscaled images.

## Use cases
- upscale low-resolution images 4x with a neural network
- implement sub-pixel convolution layers in TensorFlow
- reproduce the ESPCN super-resolution paper
- train a super-resolution model on custom images
- compare transposed convolutions vs sub-pixel convolutions
- learn how super-resolution CNNs work

## When to choose
- you need a reference implementation of sub-pixel convolution in TensorFlow
- you want to experiment with or study ESPCN-style super-resolution
- you need a lightweight starting point for image upscaling research

## When to avoid
- you need a production-ready or actively maintained super-resolution toolkit
- you use PyTorch or other frameworks instead of TensorFlow
- you need state-of-the-art super-resolution quality rather than a classic baseline

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: machine-learning, deep-learning, image-processing, computer-vision
- platform: python
- tags: super-resolution, subpixel-convolution, tensorflow, research-code, jupyter-notebook

## Member repositories
- atriumlts/subpixel (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:16.700548+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 2026-08-30T02:52:49.693678+00:00, confidence not recorded.
  - readme: https://github.com/atriumlts/subpixel (fetched 2026-08-28T04:06:16.700548+00:00, sha bd580023ec34)
- Data as of 2026-08-30T08:39:29.467469+00:00.