A neurofuzzy classification system for the effects of diabetes mellitus on ophtalmic artery

• Published in: Journal of medical systems Vol. 28; no. 2; pp. 167 - 176
• Main Authors: Serhatlioğlu, Selami, Hardalaç, Firat, Kiriş, Adem, Ozdemir, Hüseyin, Yilmaz, Turgut, Güler, Inan
• Format: Journal Article
• Language: English
• Published: United States 01-04-2004
• Subjects: Artery disease; Blood Flow Velocity; Case-Control Studies; Classification schemes; Diabetes; Diabetes Mellitus - physiopathology; Doppler signals; Fuzzy Logic; Humans; Medicine; Ophthalmic Artery - diagnostic imaging; Ophthalmic Artery - physiopathology; Turkey; Ultrasonography, Doppler
• ISSN: 0148-5598

Here, we present a classification system for the effects of diabetes mellitus (DM) on blood flow hemodynamics of the ophthalmic arteries by using neurofuzzy system. Blood flow hemodynamics were obtained from 80 ophthalmic arteries of 20 healthy persons and 20 patients with DM by using 7.5 MHz transducer and Doppler-M unit. Peak systole, peak diastole, resistive index (RI), pulsatile index (PI), and systole/diastole rate (SDR) were measured with the use of Doppler sonography. These values were applied to neurofuzzy system using NEFCLASS model. The performance of this classification system was examined with the application of the data obtained from Doppler analyses of the right and left ophthalmic arteries to the neurofuzzy system. After learning and testing processes, 85% success rates were reached from the data of right ophthalmic arteries, and 87.5% success rates were reached from the data of left ophthalmic arteries. Our findings suggest that neurofuzzy system may provide a successful classification system for the effects of DM on either right or left ophthalmic arteries with the application of Doppler signal parameters from carotid arteries to neurofuzzy system may produce a new and reliable classification system for diagnosing diameter stenosis.