LAUSR

Computer-based image analysis system enables efficient retrieval of similar images from large-size pathology database. In such a system, images are expressed based on visual content characteristics, and similarities between images are obtained by comparing the features. A pathology image is usually very huge and expressed as several layer of image quality called OpenSlide. To find similar cells from a OpenSlide file, we have to use high performance computer equiped with multi-core and large size memory. In this paper, we propose a method to find similar cell images with resource limited computer. For this purpose, we exploit several technique to minimize system resource requirement and adapt imaging process scheme that enhances the accuracy of finding similar cell images from a OpenSlide file. We adapt a leveling, tiling and sub tiling to the OpenSlide file and extracting the feature points accurately using the hybrid feature extracting algorithm that adapts advantages of ORB and Blob algorithm. Furthermore, grayscale and histogram schemes are used to improve the accuracy of finding similar pathology cell images. Experiment results show that the proposed system improves the performance of the system and increases the accuracy of finding similar images efficiently.