Authors:
L. S. Brea
1
;
N. Barreira
1
;
A. Mosquera
2
;
H. Pena-Verdeal
2
and
E. Yebra-Pimentel
2
Affiliations:
1
Univ. A Coruna, Spain
;
2
Univ. Santiago de Compostela, Spain
Keyword(s):
Image Processing, Medical Imaging, Pattern Recognition.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Computational Intelligence
;
Data Manipulation
;
Evolutionary Computing
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Machine Learning
;
Methodologies and Methods
;
Neurocomputing
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Sensor Networks
;
Soft Computing
;
Symbolic Systems
Abstract:
Hyperemia is the occurrence of redness in a certain tissue. When it takes place on the bulbar conjunctiva, it can be an early symptom of different pathologies, hence, the importance of its quick evaluation. Experts grade hyperemia as a value in a continuous scale, according to the severity level. As it is a subjective and time consuming task, its automatisation is a priority for the optometrists. To this end, several image features are computed from a video frame that shows the patient’s eye. Then, these features are transformed to the grading scale by means of machine learning techniques. In previous works, we have analysed the performance of several regression algorithms. However, since the experts only use a finite number of values in each grading scale, in this paper we analyse how classifiers perform the task in comparison to regression methods. The results show that the classification techniques usually achieve a lower training error value, but the regression approaches classif
y correctly a larger number of samples.
(More)