DSLR Imperfections Extraction from Image for Source Detection

Abstract

Digital images frequently contain the objects which are not true for the depicted scene that can completely change the human perception of provided information. This article proposes an approach to search of the digital image forensics caused by distortions and noise of the Digital Single Lens Reflex (DSLR) devices. The tampered object detection algorithms can find these distortions and imperfections despite their size. Therefore it is necessary to reduce influence of these factors in various stages of image processing and forming in the digital devices. The method of determining the reliability of test image forming in the device under consideration based on the pixel non-uniformity of the device that appears at incident light was suggested.

Keywords

• pixel non-uniformity
• DSLR imperfections
• image correlation