Data forensics analysis on biometric images using Benford's law and support vector machine algorithm

The manipulation of biometric data has become a prominent topic, leading to the development and exploration of methods for detecting such manipulation. This study utilizes a combination of Benford’s law, an image quantization to analyze the fingerprint images processing associated with biometric data. The aim is to propose a mechanism for detecting data manipulation, particularly when one biometric sample is substituted for another in an application, whether intentionally or unintentionally. The study focuses on differentiating between biometric samples and investigating the modification of fingerprint images. To achieve this, the Benford legal difference scale is applied to fingerprints digitally obtained, industrially created fingerprints, contactless acquired fingerprints to search for separation modes. Benford’s law has been successful in determining the alteration of landscape images in previous studies, and this study combines Benford’s law elements with a SVM (Support Vector Machine) to identify malicious alterations in JPEG fingerprint images. The proposed strategy is intended to safeguard against internal threats.

Text 118377.pdf Download (2MB) Official URL or Download Paper: http://ethesis.upm.edu.my/id/eprint/18373

Actions (login required)

View Item