MU Digital Repository
Logo

Source Camera Identification Using Noise Residual

Akshatha, K R and Anitha, H and Karunakar, A K and Raghavendra, U and Shetty, Dinesh (2016) Source Camera Identification Using Noise Residual. In: International conference on Recent Trends in Electronics, Information & Communication Technology, 20/05/2016, SVCE Bangalore.

[img] PDF
242.pdf - Published Version
Restricted to Registered users only

Download (539kB) | Request a copy

Abstract

Abstract—Source camera identification process plays a major role in digital image forensics, which identifies the source camera used for capturing the given image. The state-of-the-art methods for source camera identification are mainly based on identifying the demosaicing artifacts, sensor noise based artifacts or by using image features from spatial/frequency domain. This paper proposes a novel feature based approach for source camera identification, by making use of the noise patterns present in the images, caused due to the individual source cameras. The scene contents present in the images are eliminated using Gaussian based denoising to estimate the noise patterns present in the images. Higher Order Wavelet Statistics (HOWS) are used as discriminative features and are used with support vector machine (SVM) classifiers to identify the originating source camera of the given image. The experiment has been carried out using the images captured from different cell phone cameras and the obtained results have proved the robustness of the proposed method.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Digital image forensics, Source camera identification, Sensor noise, Higher order wavelet statistics, Support vector machines, Ten-fold cross validation
Subjects: Engineering > MIT Manipal > Electronics and Communication
Engineering > MIT Manipal > Instrumentation and Control
Engineering > MIT Manipal > MCA
Depositing User: MIT Library
Date Deposited: 16 Jul 2016 11:47
Last Modified: 16 Jul 2016 11:47
URI: http://eprints.manipal.edu/id/eprint/146620

Actions (login required)

View Item View Item