U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Camera Technology Survey of 3D Offerings and 2D Occlusions

NCJ Number
Date Published
December 2012
25 pages
This report describes the methodology and presents the findings of the market survey for 3D cameras and camcorders capable of face collections at 50 meters or greater, as well as an assessment of 2D camera background removal algorithms and techniques.
The National Institute of Standards and Technology (NIST) has proven that face recognition can be accurate with existing face-matching algorithms, provided the probe and dataset images are of excellent quality. Unfortunately, high-resolution, studio-quality images can rarely, if ever, be collected covertly by surveillance cameras. The current market survey served the objective of the National Institute of Justice's (NIJ's) Sensor, Surveillance, and Biometric Technologies (SSBT) Center of Excellence's effort to develop a surveillance system that can capture face images of uncooperative individuals at a distance greater than 50 meters in unconstrained environments, followed by automated matching of the collected face against mug shot datasets. The market survey sought to determine whether current technologies are capable of achieving the aspirations for such a surveillance system. The overall conclusion of the survey is that there are no identified 3D cameras or camcorders capable of collecting face images at 50 meters or greater that can provide the resolution and focal length required to provide 90 pixels between pupil centers. The 3D camcorder products reviewed in the survey as potential face-recognition image-collection systems were the Sony DEV-3, JVC GS-TD1, and Panasonic HDC-10000. Neither the JVC nor the Panasonic are ideal solutions for collecting face images at a distance. The Sony DEV-3 is currently collecting face-image data at West Virginia University as part of the evaluation of the prototype system NIJ SSBT CoE SVI binocular developed with NIJ funding. The most promising research area associated with image face segmentation is in computational photography. Appended product and research Web links and camcorder specifications