Automatic identification of ROI in figure images toward improving hybrid (text and image) biomedical document retrieval

Abstract

Biomedical images are often referenced for clinical decision support (CDS), educational purposes, and research. They appear in specialized databases or in biomedical publications and are not meaningfully retrievable using primarily textbased retrieval systems. The task of automatically finding the images in an article that are most useful for the purpose of determining relevance to a clinical situation is quite challenging. An approach is to automatically annotate images extracted from scientific publications with respect to their usefulness for CDS. As an important step toward achieving the goal, we proposed figure image analysis for localizing pointers (arrows, symbols) to extract regions of interest (ROI) that can then be used to obtain meaningful local image content. Content-based image retrieval (CBIR) techniques can then associate local image ROIs with identified biomedical concepts in figure captions for improved hybrid (text and image) retrieval of biomedical articles.In this work we present methods that make robust our previous Markov random field (MRF)-based approach for pointer recognition and ROI extraction. These include use of Active Shape Models (ASM) to overcome problems in recognizing distorted pointer shapes and a region segmentation method for ROI extraction.We measure the performance of our methods on two criteria: (i) effectiveness in recognizing pointers in images, and (ii) improved document retrieval through use of extracted ROIs. Evaluation on three test sets shows 87% accuracy in the first criterion. Further, the quality of document retrieval using local visual features and text is shown to be better than using visual features alone.

Accessibility of graphics in STEM research articles: Analysis and proposals for improvement

“…Labels inside the image such as text and symbols are another kind of textual information related to figures in academic articles. They often highlight regions of interest (ROIs) (Antani, Demner‐Fushman, Li, Srinivasan, & Thoma, ) and are frequently referenced in the captions or mentions in biomedical publications (You et al., ).…”

“…Traditionally, the function and use of image‐related textual descriptions in academic publications (especially in the area of biomedicine) have been analyzed from the perspective of improving the efficiency and satisfaction of the image retrieval process (Divoli, Wooldridge, & Hearst, ); as an element of the figure in the hybrid (text and image) biomedical document retrieval process (Apostolova et al., ; Christiansen, Lee, & Chang, ; You et al., ); and for summarizing the image content (Agarwal & Yu, ; Bhatia, Lahiri, & Mitra, ; Neveol, Deserno, Darmoni, Guld, & Aronson, ; Yu & Lee, ). Caption extraction from PDF documents in the domains of chemistry, computer science, physics, and astronomy has been studied by Choudhury et al.…”

Do Physicians Make Their Articles Readable for Their Blind or Low-Vision Patients? An Analysis of Current Image Processing Practices in Biomedical Journals from the Point of View of Accessibility

“…A growing number of clinicians, educators, researchers, and other professionals use digital images in their work [20]; biomedical images are often referenced for clinical decision support, educational purposes, and research [21], and specialized collections of biomedical images are considered of value for research and training purposes [22]. Health care professionals regularly search for images published in medical journals to find specialized information [23].…”