Philippine E-Journals

Human Ethnicity-Based Greeter: An Application of Fisherface and Bayes’ Theorem

Lourdeline G. Plaza | Marvin John B. Zabate

Discipline: Computer Science

 

Abstract:

This study was all about the tandem of Fisherface and Bayes Theorem in the identification of human ethnicity. Ethnicity recognition also belongs to facial recognition but it emphasizes on identifying human ethnicity. The identification is based on static image as inputs. The application is limited to identify two groups of ethnicity. The first group consists of ethnicities that belong to East Asian Countries. The second group consists of ethnicities that belong to South East Asian Countries. People whose nationalities are not mentioned in the first and the second group were assigned as a default group who uses English in the greeting. The identification was done with the use of the said two methods and afterwards, the application performed a greeting based on the language that the said group is using. The result of the identification may vary depending on some factors that the input image has. The use of Fisherface and Bayes Theorem in the recognition of human ethnicity can give information used in identifying the group of the input image. The researchers concluded that there are a lot of factors to be considered in the recognition of faces and that these factors have to be carefully considered because they can affect the result.