BioTamperNet: Affinity-Guided State-Space Model Detecting Tampered Biomedical Images

Soumyaroop Nandi, Prem Natarajan, [0.5ex] USC Information Sciences Institute, Marina del Rey, CA, USA · 2026

Read BioTamperNet: Affinity-Guided State-Space Model Detecting Tampered Biomedical Images by Soumyaroop Nandi, Prem Natarajan, [0.5ex] USC Information Sciences Institute, Marina del Rey, CA, USA (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-2602-01435.txt.

Description

Soumyaroop Nandi and Prem Natarajan propose BioTamperNet for spotting copied regions in doctored biomedical images. Forensic tools trained on natural images often miss on biomedical data, where small changes undermine validity. BioTamperNet leans on affinity-guided attention from State Space Models. One module captures similarities inside one image, another maps matches across images. Light linear attention keeps localization sharp yet efficient. Trained end to end, the tool flags doctored regions and source copies at once. Tests on benchmark bio-forensic sets show clear gains over baselines in finding copied regions.

Catalog record

CREATORSoumyaroop Nandi, Prem Natarajan, [0.5ex] USC Information Sciences Institute, Marina del Rey, CA, USA
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39607
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01435

Browse the full library · About the Tower of Babel