Support vector machine active learning for image retrieval

Simon Tong, Edward Chang
2001 Proceedings of the ninth ACM international conference on Multimedia - MULTIMEDIA '01  
Relevance feedback is often a critical component when designing image databases. With these databases it is di cult to specify queries directly and explicitly. Relevance feedback interactively determinines a user's desired output or query concept by asking the user whether certain proposed images are relevant or not. For a relevance feedback algorithm to be e ective, it must grasp a user's query concept accurately and quickly, while also only asking the user to label a small number of images.
more » ... number of images. We propose the use of a support vector machine active learning algorithm for conducting e ective relevance feedback for image retrieval. The algorithm selects the most informative images to query a user and quickly learns a boundary that separates the images that satisfy the user's query concept from the rest of the dataset. Experimental results show that our algorithm achieves signi cantly higher search accuracy than traditional query re nement s c hemes after just three to four rounds of relevance feedback.
doi:10.1145/500141.500159 fatcat:xar7xfy3kbayvnjn755govzp7i