Adaptive learning for relevance feedback: Application to digital mammography

Abstract

Purpose: With the rapid growing volume of images in medical databases, development of efficient image retrieval systems to retrieve relevant or similar images to a query image has become an active research area. Despite many efforts to improve the performance of techniques for accurate image retrieval, its success in biomedicine thus far has been quite limited. This article presents an adaptive content-based image retrieval ͑CBIR͒ system for improving the performance of image retrieval in mammographic databases. Methods: In this work, the authors propose a new relevance feedback approach based on incremental learning with support vector machine ͑SVM͒ regression. Also, the authors present a new local perturbation method to further improve the performance of the proposed relevance feedback system. The approaches enable efficient online learning by adapting the current trained model to changes prompted by the user's relevance feedback, avoiding the burden of retraining the CBIR system. To demonstrate the proposed image retrieval system, the authors used two mammogram data sets: A set of 76 mammograms scored based on geometrical similarity and a larger set of 200 mammograms scored by expert radiologists based on pathological findings. Results:The experimental results show that the proposed relevance feedback strategy improves the retrieval precision for both data sets while achieving high efficiency compared to offline SVM. For the data set of 200 mammograms, the authors obtained an average precision of 0.48 and an area under the precision-recall curve of 0.79. In addition, using the same database, the authors achieved a high pathology matching rate greater than 80% between the query and the top retrieved images after relevance feedback. Conclusions: Using mammographic databases, the results demonstrate that the proposed approach is more accurate than the model without using relevance feedback not only in image retrieval but also in pathology matching while maintaining its effectiveness for online relevance feedback applications.

A Novel Similarity Learning Method via Relative Comparison for Content-Based Medical Image Retrieval

“…The similarity measure with this optimal a will be utilized in the testing step for retrieval performance evaluation. Mean intensity inside the detected nucleus contour [8][9][10] Color components (HSI) inside the detected nucleus contour 11…”

“…In [9], a computer-aided expert system named MAMMO/ICON is introduced for automated mammographic image retrieval, which is one of the earliest CBMIR studies in mammography. In [10], the authors concentrate on the relevance feedback issue in CBMIR and propose a new strategy to improve the performance of mammographic image retrieval via incremental learning and support vector machine (SVM) regression. In [11], the similarity problem is tackled in CBMIR and a hierarchal learning structure, in which neural networks (NN) and SVM are utilized as classifiers, is introduced.…”

“…Following the above understanding, much research effort has been emphasized on individuals of them. For instance, [12] focuses on different types of feature extracted from images; [11] and [13] tackle the similarity measure problem; while [10] and [14] concentrate more on the user interaction (relevance feedback) problem. In this study, the similarity measure problem in CBMIR is concentrated, and a novel similarity learning method via relative comparison is presented.…”

“…There are already many CBMIR studies proposed for a wide spectrum of clinical applications in recent years. For instance, CBMIR are vastly employed in digital mammography [6][7][8][9][10][11][12][13][14][15][16]. In [9], a computer-aided expert system named MAMMO/ICON is introduced for automated mammographic image retrieval, which is one of the earliest CBMIR studies in mammography.…”

“…The performance of TBMIR based on such inaccurate textural information will certainly be badly influenced. Therefore, in order to improve the performance of medical image retrieval systems, much research effect nowadays has been put into CBMIR, which is often considered to be a good complement to the traditional TBMIR [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19].…”