Article: A texture descriptor of the camera fingerprint for source camera model and device identification Journal: International Journal of Intelligent Systems Technologies and Applications (IJISTA) 2024 Vol.22 No.2 pp.213 - 235 Abstract: Source camera identification techniques are used to recognise the origin of a digital image. Origin of a digital image refers to the particular camera model or device which captures the image. The unique identity of the acquisition device, which arises due to the photo response non-uniformity (PRNU) of the sensor, is embedded in an image. This unique identity or camera fingerprint is extracted as the noise residual of the image. In this paper, grey level co-occurence matrix of non-overlapping window local binary pattern (NOW-LBP GLCM) feature of the noise residual of an image is identified and used as the texture feature of the camera fingerprint to detect the source camera/origin of the image. This fine texture pattern of the camera sensor, is given to the multi-class SVM classifier to recognise: 1) model; 2) individual device, of the digital camera/smartphone. From the experiments conducted it is found that this texture feature can effectively distinguish the camera model from same brand cameras and the particular camera device from same camera models. This work achieves a good identification accuracy for camera model identification and camera device identification as 96.73% and 88.52% for 12 camera models and 15 camera devices respectively. Inderscience Publishers - linking academia, business and industry through research

Title: A texture descriptor of the camera fingerprint for source camera model and device identification

Authors: Aiswariya Raj; Deepa Sankar

Addresses: Division of Electronics and Communication Engineering, School of Engineering, Cochin University of Science and Technology, Cochin, Kerala, India ' Division of Electronics and Communication Engineering, School of Engineering, Cochin University of Science and Technology, Cochin, Kerala, India

Abstract: Source camera identification techniques are used to recognise the origin of a digital image. Origin of a digital image refers to the particular camera model or device which captures the image. The unique identity of the acquisition device, which arises due to the photo response non-uniformity (PRNU) of the sensor, is embedded in an image. This unique identity or camera fingerprint is extracted as the noise residual of the image. In this paper, grey level co-occurence matrix of non-overlapping window local binary pattern (NOW-LBP GLCM) feature of the noise residual of an image is identified and used as the texture feature of the camera fingerprint to detect the source camera/origin of the image. This fine texture pattern of the camera sensor, is given to the multi-class SVM classifier to recognise: 1) model; 2) individual device, of the digital camera/smartphone. From the experiments conducted it is found that this texture feature can effectively distinguish the camera model from same brand cameras and the particular camera device from same camera models. This work achieves a good identification accuracy for camera model identification and camera device identification as 96.73% and 88.52% for 12 camera models and 15 camera devices respectively.

Keywords: source camera identification; SCI; digital image forensics; camera finger print; image classification; texture feature extraction.

DOI: 10.1504/IJISTA.2024.139815

International Journal of Intelligent Systems Technologies and Applications, 2024 Vol.22 No.2, pp.213 - 235

Received: 01 Jul 2023
Accepted: 02 Nov 2023
Published online: 05 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: A texture descriptor of the camera fingerprint for source camera model and device identification

Keep up-to-date