Leveraging user query log: toward improving image data clustering

Image clustering is useful in many retrieval and classification applications. The main goal of image clustering is to partition a given dataset into salient clusters such that the images in each cluster appear visually similar to each other compared with those in other clusters. In this paper, we propose a semi-supervised clustering algorithm, which leverages the accumulated user query log to guide the clustering process. Guided by the log file, our method arranges images into small groups and constructs a graph that captures the dissimilar relations between the groups. Each group is assigned to a feasible cluster. Our analysis reveals that the probability of image points being assigned to the correct clusters is much higher by our new proposal, compared to conventional methods. Our algorithm can produce image clusters close to the ground truth and satisfying the semantic relations between the images inferred from the query log. Experimental results further confirm the superiority of ...

• Type: Conference
• Year: 2008
• Where: CIVR
• Authors: Hao Cheng, Kien A. Hua, Khanh Vu
• Keywords: CIVR 2008, Image Analysis, Image Clustering, Query Log, Semi-supervised Clustering