Why Cryptographic Hashes Fail on Images
Cryptographic algorithms like SHA-256 exhibit the avalanche effect: changing a single bit or saving a JPEG at 79% instead of 80% produces a completely unrecognizable new hash.
Perceptual hashing models human visual perception by downsampling the image and computing low-frequency DCT coefficients. If an image is resized, compressed, or cropped slightly, its pHash remains virtually identical.
Key Investigative Takeaway
Use SHA-256 for chain of custody and legal integrity; use pHash for tracking duplicate image distribution across social media.