Digital Camera Identification

• NCJ Number: 196837
• Journal: Journal of Forensic Identification Volume: 52 Issue: 5 Dated: September/October 2002 Pages: 621-632
• Author(s): Zeno J. Geradts; Jurrien Bijhold; Martijn Kieft; Kenji Kurosawa; Kenro Kuroki; Naoki Saitoh
• Date Published: September 2002
• Length: 12 pages
• Publication Format: Article
• Publication Type: Report (Study/Research)
• Language: English
• Country: United States of America

Annotation

In an effort to establish a nexus between a specific photograph and a specific digital camera, this research examined defects in the charge-coupled device (CCD), noise introduced by the pixel arrays, file formats, and manufacturer watermarkings.

Abstract

The study examined various CCD cameras, from economy 640 x 480 low resolution to higher resolution, 2-million pixel cameras. When CCD's are manufactured, defects sometimes occur. The defects are often noted in the manufacturer's data sheets. CCD's with a large number of defects are less expensive than CCD's with few defects. In addition to pixel defects that occur during the manufacturing process, new pixel defects may occur during the lifetime of the CCD. This article briefly describes some common CCD defects among some manufacturers. Twelve economy cameras of the same model were tested, and pixel defects were detected by comparing five images acquired by each camera and identifying the pixel defects that were reproducible in these images. Each camera had at least five defective pixels, and each had pixel defects in different places on the CCD. Because temperature can affect the performance of the CCD, the cameras were tested at different temperatures. The more expensive cameras have better CCD's with fewer errors. These errors were investigated by averaging at least 10 images; however, the errors were not visible in the Sony Cybershot, Sony Mavica, Sony FD83, or Sony Handycam. Since compression of the images might influence the visibility of pixel defects, tests were conducted on the cameras with detectable defects (the economy cameras). Apparently the position of the pixel did not change until a compression of 50:1 was used. Regarding the comparison of files, a hex editor can be used to examine image headers. Image header information can be compared for type of information included and layout within the header. Comparisons of noise levels are possible by making an image of the same object and comparing the images with one another. There were no watermarks observed in the examined cameras. 6 figures and 13 references