Comparing Machine Learning Techniques in a Hyperemia Grading Framework

• DOI: 10.5220/0005756004230429
• Publication Date: 2016-04-28T05:19:28Z
• Authors: L. S. Brea; N. Barreira; A. Mosquera; H. Pena-Verdeal; E. Yebra-Pimentel
• Keywords: Grading (engineering); Grading scale

Hyperemia is the occurrence of redness in a certain tissue. When it takes place on bulbar conjunctiva, can be an early symptom different pathologies, hence, importance its quick evaluation. Experts grade hyperemia as value continuous scale, according to severity level. As subjective and time consuming task, automatisation priority for optometrists. To this end, several image features are computed from video frame that shows patient’s eye. Then, these transformed grading scale by means machine learning techniques. In previous works, we have analysed performance regression algorithms. However, since experts only use finite number values each paper analyse how classifiers perform task comparison methods. The results show classification techniques usually achieve lower training error value, but approaches classify correctly larger samples.