1 January 2004 Characterization of artifacts in fully digital image-acquisition systems: Application to web cameras
Author Affiliations +
Optical Engineering, 43(1), (2004). doi:10.1117/1.1626667
We apply principal component analysis (PCA) to the characterization of artifacts in a digital image-acquisition system containing image-compression algorithms. The method is successfully applied to web cameras. The classification done with the PCA method produces three processes. The pure spatial process retrieves the luminance distribution of a static object. The pure temporal process is directly related with the temporal noise of the system. An intermediate spatial-temporal process reveals the interaction between the compression algorithms and the spatial-frequency contents of the object. Without prior information, the PCA method is able to distinguish this interaction from the classical temporal noise. The analysis of the anomalous pixels also reveals the location in the scene where the compression algorithms work harder. An extension of this analysis identifies the origin of the anomalous behavior in terms of its spatial or temporal character.
José Manuel López-Alonso, Javier Alda, "Characterization of artifacts in fully digital image-acquisition systems: Application to web cameras," Optical Engineering 43(1), (1 January 2004). http://dx.doi.org/10.1117/1.1626667

Principal component analysis


Spatial frequencies

Image compression


Analog electronics

Optical engineering


Multifunctional Macromolecules
Proceedings of SPIE (May 03 1988)
Fingerprint ridge structure generation models
Proceedings of SPIE (February 01 1998)
Real-time enhanced vision system
Proceedings of SPIE (May 25 2005)
Efficient disparity vector coding for multi-view 3D displays
Proceedings of SPIE (February 04 2010)

Back to Top