NCJ Number
193727
Journal
Journal of Forensic Sciences Volume: 47 Issue: 1 Dated: January 2002 Pages: 97-102
Date Published
2002
Length
6 pages
Annotation
This paper examines the method that can be employed to determine whether or not questioned video images were recorded with a specific video camera.
Abstract
Recorded voices are often used and images are often analyzed for individual human identification and reconstruction of criminal scenes. This type of identification is similar to the traditional identification or word processor models by printouts. However, the method presented in this paper may be used as an individualization technique, but cannot be employed as a categorization or model estimation technique. While this kind of individualization method is useful for examining video-recorded images of kidnapping or child pornography, or for demonstrating the evidence of video editing, it has not been widely used. The investigators examined the electrical properties of charge coupled devices (CCDs) for individualizing video cameras. The approach is based on the nonhomogeneous nature of dark currents in CCDs. The distribution pattern of the nonhomogeneous dark currents is unique and intrinsic to a specific camera. The distribution patterns of dark currents in nine cameras representing four different types were examined. The results showed that it was possible to identify if a given image was recorded with a specific camera. In eight of the nine cameras three types of unique detectable patterns were identified in recorded blank images, demonstrating that it was possible to know whether or not a given image had been recorded with a certain camera. In addition, the method employed was used in an actual case to determine if the questioned video images of a criminal scene were recorded with the suspect’s camera. The subsequent analysis showed that the video image had been recorded with the suspect’s camera because coordinates of three hot pixels on the images and the camera were equal. 10 References