With the ever-growing volume of images used in medicine, the capability to retrieve relevant images from large databases is becoming increasingly important. Despite the recent progress made in the field, its applications in Computer-Aided Diagnosis (CAD) thus far have been limited by the ability to determine the intrinsic mapping between high-level user perception and the underlying low-level image features. Relevance Feedback (RFB) is a post-query process to refine the search by using positive and/or negative indications from the user about the relevance of retrieved images, which has been applied successfully in traditional text-retrieval systems for improving the results of a retrieval strategy. In this chapter, the authors review some recent advances in RFB technology, and discuss its expanding role in content-based image retrieval from medical archives. They provide working examples, based on their experience, for developing machine-learning methods for RFB in mammography and highlight the potential opportunities in this field for CAD applications and clinical decision-making.
TopIntroduction
With the growing volume of images used in medicine, the capability to retrieve relevant images from large databases is becoming increasingly important (Bimbo, 1999; Rui & Huang, 1999). The key to a successful image retrieval system lies in the development of appropriate similarity metrics for ranking the relevance of images in a database to the query image. Content-Based Image Retrieval (CBIR) has been proposed to overcome the difficulties encountered in textual annotation for large image databases. Relevance Feedback (RFB) is a post-query process to refine the search by using positive and/or negative indications from the user of the relevance of retrieved images. It has been applied successfully in traditional text-retrieval systems for improving the results of a retrieval strategy (Baeza-Yates & Ribeiro-Neto, 2011; Chowdhury, 2010; El-Naqa, Yang, Galatsanos, & Wernick, 2002).
Despite the progress made in the general area of image retrieval in recent years, its success in biomedicine has been quite limited until recently, where significant evolution has taken place (El Naqa, Wei, & Yang, 2010; Muller, Michoux, Bandon, & Geissbuhler, 2004; Shapiro, et al., 2008; Wong, 1998). In our previous work, we investigated the use of CBIR for digital mammograms (El Naqa, Yang, Galatsanos, Nishikawa, & Wernick, 2004; El-Naqa, et al., 2002). The goal was to provide the radiologist with a set of images from past cases that are relevant to the one being evaluated, along with the known pathology of these cases. We conjecture that by presenting images with known pathology that are “visually similar” to the image being evaluated, a mammogram retrieval system may serve as a more intuitive aid to radiologists, potentially leading to improvement in their diagnostic accuracy. Furthermore, it is expected that such a technique would be a useful aid in the training of students and residents, since it would allow them to view images of lesions that appear similar, but may have differing pathology.
In this chapter, we review some recent advances in RFB technology, and discuss its expanding role in CBIR from biomedical archives. We provide working examples based on our experience in applying different RFB strategies in mammography with a special focus on developing machine-learning methods to improve effectiveness and efficiency of CBIR for mammography. Finally, we highlight the current challenges and the potential opportunities in this field for CAD applications and computerized clinical decision-making.