abstract

• In recent years, the search of efficient and noninvasive methods for the diagnosis of diseases has grown among the scientific community. One of the explored areas for that purpose is the analysis of the pupillary response to light stimulus, obtaining results in areas such Diabetes, Alzheimer, Neurological Disorders, Melancholia and other different physiological states. However, each of those investigations are focused on a concrete pathology or case of study and the data is analyzed in different ways, according the pathology that is searched. For that reason, we proposed a first scope to, through clustering techniques, find a model to be able to detect different groups of individuals with characteristics in common using the pupillary response time. The experiment was done over data coming from 32 healthy individuals and a characterization of their pupillary response, then a clustering algorithm was employed to group the data and finally an analysis of the clusters was performed. The result gives four clusters of which two of them have a high percentage of individuals with overweight. Thus it concludes that is possible to identify groups of people with characteristics in common using the pupillary response time as principal indicator and as future work is proposed a deep study using individuals with specific diagnosed pathologies and healthy individuals. © 2017 IEEE.

authors

• Guevara Codina Edgar
• Hernández G.

publication date

• 2017-01-01

published in

• 2017 International Conference on Electronics, Communications and Computers, CONIELECOMP 2017  Proceedings

keywords

• Clustering techniques; K-means; Overweight; Pupillary response time; Pupillometry Cluster analysis; Diagnosis; Noninvasive medical procedures; Pathology; Physiological models; Clustering techniques; K-means; Overweight; Pupillary response; Pupillometry; Clustering algorithms

Digital Object Identifier (DOI)

• 10.1109/CONIELECOMP.2017.7891817

PubMed ID

start page

end page

volume

issue