Classification of retinal vessels using a collaborative agent-based architecture

Article type: Research Article

Authors: Chamoso, Pablo^{a; *} | Rodríguez, Sara^{a; **} | de la Prieta, Fernando^a | Bajo, Javier^b

Affiliations: [a] Department of Computer Science, University of Salamanca, IBSAL/BISITE Research Group. Edificio I+D+I, 37007, Salamanca, Spain. E-mails: chamoso@usal.es, srg@usal.es, fer@usal.es | [b] Department of Artificial Intelligence, Technical University of Madrid, Campus Montegancedo, Boadill a del Monte, 28660, Madrid, Spain. E-mail: jbajo@fi.upm.es

Correspondence: [*] Corresponding author. E-mail: chamoso@usal.es.

Correspondence: [**] Corresponding author. E-mail: srg@usal.es.

Abstract: Recent researches have proved a direct relationship between data gathered from retinal images and some pathologies such as hypertension or diabetes. There are a lot of studies that propose methods used in this particular area of investigation. They are based on image processing algorithms to extract retinal information, but there is no general accepted methodology to obtain the information automatically. In this work, a new methodology for the classification of retinal vessels based on agreement technologies and artificial vision techniques is presented. The proposed methodology has been evaluated by expert users and compared with other existing free software with similar features.

Keywords: Agents, retinal vessels, agreement technologies, visual analysis, e-health, virtual organizations

DOI: 10.3233/AIC-180772

Journal: AI Communications, vol. 31, no. 5, pp. 427-444, 2018