A CNN Based Algorithm for Medical Images Correlation

• Published in: Journal of Computer Science and Control Systems Vol. 4; no. 2; p. 69
• Main Authors: Tepelea, Laviniu, Gavrilut, Ioan, Gacsádi, Alexandru
• Format: Journal Article
• Language: English
• Published: Oradea University of Oradea 01.10.2011
• Subjects: Algorithms; Computation; Correlation; Correlation analysis; Diagnosis; Field programmable gate arrays; Hardware; Medical; Neural networks; Studies
• ISSN: 1844-6043; 2067-2101

This paper propose a CNN algorithm to process medical images for diagnosis. The algorithm is based on image correlation in CNN manner. Image correlation generally needs a lot of time to get a result. In CNN domain most of operations are achieved by parallel processing, so we can compute the correlation coefficients between two images in parallel manner. We can use such algorithm in a medical diagnosis assistance system, to process and analyze computer tomography images or in a system developed to assist people with visual impairments in moving. An analog chip for CNN computing is very expensive, but an FPGA hardware is not so expensive and is more popular. Now we can implement the correlation algorithm in CNN manner on FPGA. [PUBLICATION ABSTRACT]