Trust Stamp gains approval for a new patent around ‘lossy’ biometric representation

Trust Stamp has gained approval for a new patent from the United States Patent and Trademark Office, pertaining to ‘lossy’ biometric representation.

The newly allowed patent, dubbed “Systems and processes for lossy biometric representation,” concerns the company’s biometric hashing technology, which it uses for subject identity verification.

As per the announcement, this process uses “neural network processing and pseudorandom matrix multiplication to create anonymized vector representations of biometric data that can be compared to confirm a subject’s identity.”

These anonymized vector representations are ‘lossy’ meaning, according to the company’s claims, not all information from the original biometric is included in the anonymized vector representation.

These ‘lossy’ representations will purportedly allow for enhanced security when dealing with a subject’s sensitive biometric data, while still allowing for the anonymized vector representations to be compared to one another for the purposes of identity verification.

The process from the patent has already been integrated into the company’s “Irreversibly Transformed Identity Token (IT2)” technology, a solution that looks to replace the storage and use of biometric templates with an “irreversibly transformed identity token, or IT²” which it says is generated by AI.

The news comes as the Atlanta, Georgia -based firm has been active in terms of pushing through new patents during 2023. The firm received a Notice of Allowance for three different patents last month, relating to securing biometric data and as well as the use of biometric tokens in the metaverse.

According to Dr. Norman Poh, Trust Stamp’s Chief Science Officer, his company has “30 patents now issued, allowed, or pending, over the last seven years.”

Article Topics

biometric hash  |  biometrics  |  patents  |  tokenization  |  Trust Stamp