Duplicate Photo Hunter

01Overview

Duplicate Photo Hunter scans a folder for both byte-identical and visually similar photos, groups them with previews, and highlights the best copy to keep. Its default behaviour is to delete nothing: selected duplicates are moved to a backup folder.

There are two detection modes. Exact duplicates are found from file bytes via SHA-256; visual similarity is computed with a DCT-based perceptual hash (pHash) and Hamming distance, which also catches resized, re-compressed or lightly edited copies.

02Key Features

• Two detection modes — SHA-256 for exact duplicates, pHash + Hamming for visual similarity
• Adjustable threshold — a live Strict → Balanced → Loose slider
• Non-freezing UI — scanning runs on a background thread with a live progress bar
• Rich result cards — thumbnail, file name, resolution, size and full path
• Smart keep suggestion — highest resolution → largest file → newest date
• Safe cleanup — duplicates move to an auto-created, collision-safe Duplicates_Backup folder
• Instant re-grouping — changing mode or threshold updates results without re-reading files

03Engineering Notes

• Hashes are computed once and kept in memory; changing the threshold only re-groups the results instead of re-reading the disk.
• A before/after summary (photos scanned, duplicate groups, space recoverable) shows the impact of the operation up front.
• Written in C++17 with Qt 6 Widgets and built with CMake; the dark theme is tuned to stay readable both on screen and in screen recordings.