A proposal based on color descriptors and local binary patterns histogram as support tool in presumptive diagnosis of hiatus hernia

Research output: Contribution to conferencePaper

1 Scopus citations

Abstract

© 2014 IEEE. According to the World Health Organization (WHO), as of 2012 esophageal cancer is the eighth-most common cancer globally with 456,000 new cases during the year. One of the triggers of the esophageal cancer is the hiatus hernia, and currently, the frequency disease increases with age, from 10% in patients younger than 40 years to 70% in patients older than 70 years old. Given the above, the aim of this paper is to provide a support tool in the presumptive diagnosis of the hiatus hernia. In the proposed approach we have used two kind of descriptors to provide hiatus hernia's presumptive diagnosis: a first one based on color spaces (RGB, HSV, CIELab and CIELuv) and a second one based on texture descriptors (Local Binary Pattern Histograms). During the experiments, we have tested two kinds of classifiers: K Nearest Neighbor and Random Forest, on a corpus of 48 real cases of images of healthy patients and patients suffering from hiatus hernia. The results are promising, achieving 83% accuracy in classifying the disease.
Translated title of the contributionUna propuesta basada en descriptores de color e histograma de patrones binarios locales como herramienta de apoyo en el diagnóstico presuntivo de hernia de hiato.
Original languageEnglish
DOIs
StatePublished - 9 Feb 2014
Event2014 IEEE International Autumn Meeting on Power, Electronics and Computing, ROPEC 2014 - Ixtapa, Mexico
Duration: 5 Nov 20147 Nov 2014

Conference

Conference2014 IEEE International Autumn Meeting on Power, Electronics and Computing, ROPEC 2014
Abbreviated titleROPEC 2014
Country/TerritoryMexico
CityIxtapa
Period5/11/147/11/14

Fingerprint

Dive into the research topics of 'A proposal based on color descriptors and local binary patterns histogram as support tool in presumptive diagnosis of hiatus hernia'. Together they form a unique fingerprint.

Cite this