CFE-PPAR: Compression-friendly encryption for privacy-preserving action recognition leveraging video transformers

2026

Read CFE-PPAR: Compression-friendly encryption for privacy-preserving action recognition leveraging video transformers (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-05692.txt.

Description

Researchers at Chiba University's Graduate School of Informatics tackle a privacy flaw in encrypted video recognition. This task lets machines classify human activity in video without exposing sensitive visual content. Encryption-based methods protect privacy well and keep accuracy high. But they break down when encrypted videos are compressed, and both recognition and visual quality collapse. The team's method, CFE-PPAR, is the first compression-friendly encryption approach for this task.

Videos encrypted with secret keys can be read directly by a video transformer whose settings are transformed with the same keys. Tests on the UCF101 and HMDB51 action sets under Motion-JPEG and H.264 compression show CFE-PPAR beating earlier methods. Authorized clients can still decrypt content for forensic use, so the system guards privacy in transit and storage without surrendering recognition accuracy.

Catalog record

YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.05692

Browse the full library · About the Tower of Babel