Image feature selection is an important issue for source camera identification. Well-selected features should make camera classifiers accurate, efficient as well as robust. Current source camera identification schemes select image features mainly based on classification accuracy and computational efficiency. In this work, we demonstrate that robustness should also be considered for classifiers which aim at real-world tasks. Besides, we reveal what impact the reduced feature subset will have on the robustness of camera classifiers. The dimensionality reduction is often necessary for computational efficiency.
|Title of host publication||Proceedings of the Sixth International Workshop on Digital Forensics and Incident Analysis|
|Subtitle of host publication||WDFIA 2011|
|Place of Publication||UK|
|Publisher||University of Plymouth|
|Number of pages||10|
|Publication status||Published - 2011|
|Event||6th Annual Workshop on Digital Forensics & Incident Analysis: WDFIA 2011 - Kingston University, London, United Kingdom|
Duration: 07 Jul 2011 → 08 Jul 2011
|Conference||6th Annual Workshop on Digital Forensics & Incident Analysis|
|Period||07/07/11 → 08/07/11|
Hu, Y., Li, C-T., Zhou, C., & Lin, X. (2011). Issues on Selecting Image Features for Robust Source Camera Identification. In Proceedings of the Sixth International Workshop on Digital Forensics and Incident Analysis: WDFIA 2011 (pp. 1-10). University of Plymouth.