User-Assisted Image Classification on Personal Photo Collections

Image classification on personal photo collections can be extremely useful to various management tasks. However, there is little progress made forwarding it due to (1) lack of training data, and (2) subjectivity inherent in a user’s photo-organizing behavior. In this paper, we propose a framework, User-assisted image Classification on Personal Photo Collections (UCP), to address this problem. The uniqueness of this framework is that it is user-centric and includes users in the loop. Our experimental results show that the techniques used in this framework are promising.

• Type: Conference
• Year: 2006
• Where: ICMCS
• Authors: Wei-Ta Chen, Ying-Hsiang Wen, Ming-Syan Chen
• Keywords: ICMCS 2006, Image Classification, Personal Photo Collections, Various Management Tasks