Haar Wavelet Hash
Haar wavelet hashing applies the Haar wavelet transform to extract low-frequency features of an image/signal before hashing, used for perceptual hashing and image similarity comparison.
Purpose
Provides a perceptual hash robust to scaling and compression for image retrieval; similar images yield similar hashes.
Mechanism
The Haar wavelet transform is applied to the image; low-frequency coefficients (capturing overall structure) are quantized into a binary bit string as the perceptual hash. Similar images have close low-frequency coefficients, so their hashes are at small Hamming distance.
Use cases
- Perceptual image hashing
- Image similarity retrieval
- Copyright detection