Image Hashing and Comparison Project

Principal Investigator: James Goldman

The goals of the Image Hashing and Comparison Project are to develop methods for creating “fuzzy” hashes of digital images. Unlike a cryptographic digital hash which is sensitive to any change in the source content, a fuzzy hash will be tolerant of manipulations to a digital image. This allows derivative images, which have been resized, cropped, or rotated, to be linked to the original source image. A provisional patent application has been issued on the project’s accomplishments so far, and research continues on new image hashing techniques. A working prototype leading to possible product commercialization is being developed.

Personnel

  • Asawaree Kulkarni
  • Hina Chaudry

Keywords: digital images, hash, cryptography