Adversarial Shadows in Digital Forensics: New Insights Into File Fragment Classification Vulnerabilities and Defenses

Title

Adversarial Shadows in Digital Forensics: New Insights Into File Fragment Classification Vulnerabilities and Defenses

Creator

Mary, Teena; Sreeja, C.S.

Description

The paper is a comprehensive survey of adversarial attacks on file fragment classification (FFC) models - a relatively unexplored area in digital forensics, given the increasing application of machine learning techniques. Unlike image or text classification adversarial attacks, adversarial attacks on FFC exploit statistical and structural properties at the byte level in systems that lack semantic or perceptual knowledge. Such properties necessitate the use of domain-specific defense strategies, as the defense strategies adopted from other domains are typically not effective for the problems of FFC. The survey comprehensively evaluates attack mechanisms relevant to FFC, including evasion and poisoning attacks, and discusses their impact on forensic reliability. It highlights the absence of domain-specific benchmarks, robust evaluation protocols, and systematic research on the adversarial robustness of FFC. The paper also discusses the different types of byte level perturbations that can happen in fragment data, and it sets specific research priorities for raising the reliability of machine learning-based digital evidence recovery and security. The paper provides building blocks for future work, offering practical insights for development in ensuring file fragment classification systems utilized in forensics are secure. 2013 IEEE.

Source

IEEE Access;Volume;14;pp.11064-11083

Date

01-01-2026

Publisher

Institute of Electrical and Electronics Engineers Inc.

Subject

Adversarial attacks; byte level perturbations; evasion attacks; file fragment classification; forensic applications

Coverage

Mary T., CHRIST (Deemed to be University), Department of Computer Science, Karnataka, Bengaluru, 560029, India; Sreeja C.S., CHRIST (Deemed to be University), Department of Computer Science, Karnataka, Bengaluru, 560029, India, CHRIST (Deemed to be University), Center for Quantum Technologies and Complex Systems (CQTCS), Karnataka, Bengaluru, 560029, India

Rights

All Open Access; Gold Open Access

Relation

ISSN: 21693536;

Format

online

Language

English

Type

Article

Identifier

https://doi.org/10.1109/ACCESS.2026.3655822

https://www.scopus.com/pages/publications/105028406454?origin=resultslist