Spatial-temporal characterization of noise in web cameras

Web cameras are widely used as an image acquisition systems that are cheap, available and easily configured. Their settings are implemented electronically and the software drivers have a limited access to the internal characteristic of the cameras. These settings are prepared to cover the most common of the illumination conditions and scenes. The images provided by these cameras not only have to do with the physical characteristics of the detector array, but also with the compression algorithm used to present the images onto a display. In this contribution we apply the principal component analysis to a set of frames obtained by web cameras. This method allows to extract different spatial-temporal patterns of the noise of the cameras.