Content Based Radiographic Images Indexing and Retrieval Using Pattern Orientation Histogram

Abstract

Introduction: Content Based Image Retrieval (CBIR) is a method of image searching and retrieval in a <br/>database. In medical applications, CBIR is a tool used by physicians to compare the previous and current <br/>medical images associated with patients pathological conditions. As the volume of pictorial information <br/>stored in medical image databases is in progress, efficient image indexing and retrieval is increasingly <br/>becoming a necessity. <br/>Materials and Methods: This paper presents a new content based radiographic image retrieval approach <br/>based on histogram of pattern orientations, namely pattern orientation histogram (POH). POH represents <br/>the spatial distribution of five different pattern orientations: vertical, horizontal, diagonal down/left, <br/>diagonal down/right and non-orientation. In this method, a given image is first divided into image-blocks <br/>and the frequency of each type of pattern is determined in each image-block. Then, local pattern <br/>histograms for each of these image-blocks are computed. <br/>Results: The method was compared to two well known texture-based image retrieval methods: Tamura <br/>and Edge Histogram Descriptors (EHD) in MPEG-7 standard. Experimental results based on 10000 <br/>IRMA radiography image dataset, demonstrate that POH provides better precision and recall rates <br/>compared to Tamura and EHD. For some images, the recall and precision rates obtained by POH are, <br/>respectively, 48% and 18% better than the best of the two above mentioned methods. <br/>Discussion and Conclusion: Since we exploit the absolute location of the pattern in the image as well as <br/>its global composition, the proposed matching method can retrieve semantically similar medical images.