phash-similarity

A perceptual hash is a fingerprint of a multimedia file derived from various features from its content. Unlike cryptographic hash functions which rely on the avalanche effect of small changes in input leading to drastic changes in the output, perceptual hashes are "close" to one another if the features are similar.

Phash ruby implementation, compatible with Phashion

Finds the images in the directory that are most similar to the others and deletes the N most similar ones. Use to remove similar images before training stable diffusion models.

Android librarry (kotlin) : Image (JPEG, BMP) comparison (perceptual hash algorithm)

Final Year Project – AI-powered mobile app that automatically shares group photos to relevant users using face recognition, filters out duplicates and blurred images, and generates smart highlights. Built with Flutter, Node.js, AWS Lambda, PostgreSQL, and FaceNet.

[phdupes] Rust duplicate image detector using standard pdqhash implementation, GUI (egui) with 10bit display and HDR support for viewing and interacting with found duplicates, multithreaded scanning and (also RAW) image processing, encrypted metadata caching, Exif tag search, slideshow with effects, 3D LUT, GPS Map, Sun Position info

Computer Vision: Searching Similar Images powered by PHP and pHashes

An Image Deduplicator (using CNN/PHash classification models) with AVIF support

A parallelised perceptual-hash based CBIR system with OAuth2+WebAuthn security. Co-authored with @redeye-fang

FGIC - World Architectural Buildings Dataset for Multi‑Class Image Classification