Statistical significance and cognitive relevance: extraction and interpretation of discriminant visual features in face images

• Grant number: 12/22377-6
• Support Opportunities: Regular Research Grants
• Duration: March 01, 2014 - February 29, 2016
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Carlos Eduardo Thomaz
• Grantee: Carlos Eduardo Thomaz
• Host Institution: Campus de São Bernardo do Campo. Centro Universitário da FEI (UNIFEI). Fundação Educacional Inaciana Padre Sabóia de Medeiros (FEI). São Bernardo do Campo , SP, Brazil

Abstract

Similarities between biometric signals of facial images, represented by shades of pixels, geometric proportions and linear and non-linear deformations of spatial normalization of patterns, can be described as a high dimensional and sparse problem well adressed by us humans but with non-trivial scientific issues related to feature extraction and automatic coding of relevant information, classification and prediction of patterns, modeling and visual reconstruction of discriminant subspaces. Such issues are, in fact, multidisciplinary and inherent to several applications in Engineering, Computer Science and Neuroscience, among other research areas. The aim of this research project is to study the interplay between low-level visual attributes, such as color, shape and texture, and high level visual attributes, represented by semantic concepts of human reasoning, to extracting and interpreting the most discriminant features in face image analysis. The high level visual attributes are described by some supervised information like gender, age and facial expression, available on training samples and quantified by either statistical significant differences explicitly calculated from the data or cognitive relevant associations expressed implicitly by human visual perception. We expect as result of this study the implementation of a novel statistical pattern recognition method that might be a valid alternative to multivariate discriminant analysis, because it would provide a more flexible form of data compression and extract relevant features in low dimension spaces allowing better understanding and interpretation of the data for any specific a priori information of interest. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

XAVIER, IGOR R. R.; GIRALDI, GILSON A.; GIBSON, STUART JAMES; GATTAS, GILKA J. F.; RUECKERT, DANIEL; THOMAZ, CARLOS E.. Age-related craniofacial differences based on spatio-temporal face image atlases. IET IMAGE PROCESSING, v. 13, n. 9, p. 1561-1568, JUL 18 2019. (12/22377-6)