# AsuharietYgvar/AppleNeuralHash2ONNX

Convert Apple NeuralHash model for CSAM Detection to ONNX.

Repository: https://github.com/AsuharietYgvar/AppleNeuralHash2ONNX
Canonical: https://ross.abutalabs.com/products/appleneuralhash2onnx
Language: Python
License: Apache-2.0
License Family: permissive
Topics: neuralhash, apple, machine-learning, reverse-engineering
Last push: 2021-08-21T07:27:26+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": 1844, "days_push": 1838, "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 1534, forks 124 (observed 2026-08-28T04:04:59.690631+00:00)

## What it is
A Python tool that converts Apple's NeuralHash perceptual hashing model (used in Apple's CSAM Detection system) into ONNX format. It includes a demo script for computing 96-bit NeuralHash hashes of images and instructions for extracting the model files from macOS or iOS.

## Use cases
- convert Apple NeuralHash model to ONNX
- compute NeuralHash perceptual hashes for images
- reverse engineer Apple's CSAM detection hashing model
- run NeuralHash inference outside Apple platforms
- analyze how NeuralHash tolerates image resize and compression
- extract NeuralHash model files from macOS or iOS IPSW

## When to choose
- you need to run or study Apple's NeuralHash model on non-Apple platforms
- you want to independently audit or research Apple's CSAM detection hashing
- you need perceptual image hashes compatible with Apple's NeuralHash

## When to avoid
- you need a general-purpose perceptual hash like pHash and don't require NeuralHash compatibility
- you cannot obtain the proprietary model files from macOS/iOS
- you need a maintained production tool rather than a research artifact

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, reverse-engineering, image-processing, parser
- domain: machine-learning, reverse-engineering, image-processing, security
- platform: python
- tags: neuralhash, onnx, perceptual-hashing, coreml, csam-detection, model-conversion, macos, linux

## Member repositories
- AsuharietYgvar/AppleNeuralHash2ONNX (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:59.690631+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 2026-08-30T04:31:10.891535+00:00, confidence not recorded.
  - readme: https://github.com/AsuharietYgvar/AppleNeuralHash2ONNX (fetched 2026-08-28T04:04:59.690631+00:00, sha ea0240aca77c)
- Data as of 2026-08-30T08:39:29.467469+00:00.