No abstract

The aim of this work is to carry out a comparative study of face recognition methods that are suitable to work in unconstrained environments. The analyzed methods are selected by considering their performance in former comparative studies, in addition to be real-time, to require just one image per person, and to be fully online. In the study two local-matching methods, histograms of LBP features and Gabor Jet descriptors, one holistic method, generalized PCA, and two image-matching methods, SIFTbased and ERCF-based, are analyzed. The methods are compared using the FERET, LFW, UCHFaceHRI, and FRGC databases, which allows evaluating them in real-world conditions that include variations in scale, pose, lighting, focus, resolution, facial expression, accessories, makeup, occlusions, background and photographic quality. Main conclusions of this study are: there is a large dependence of the methods on the amount of face and background information that is included in the face's images, and the performance of all methods decreases largely with outdoor-illumination. The analyzed methods are robust to inaccurate alignment, face occlusions, and variations in expressions, to a large degree. LBP-based methods are an excellent election if we need real-time operation as well as high recognition rates.

Artículo de publicación ISIA unified learning framework for object detection and classification using nested cascades of boosted classifiersThis research was funded by Millenium Nucleus Center for Web Research, Grant P04-067-F, Chile. Part of the research in this paper use the FERET database of facial images collected under the FERET program. During part of this research work the authors took part of the Alfa project N◦AML//19.0902/97/06660/II-0366-FA and they would like to acknowledge its support

In this paper a robust system for enabling robots to detect and identify humans in domestic environments is proposed. Robust human detection is achieved through the use of thermal and visual information sources that are integrated to detect human-candidate objects, which are further processed in order to verify the presence of humans and their identity using face information in the thermal and visual spectrums. Face detection is used to verify the presence of humans, and face recognition to identify them. Active vision mechanisms are employed in order to improve the relative pose of a candidate object/person in case direct identification is not possible. The response of the different modules is characterized, and the proposed system is validated using image databases of real domestic environments, and human detection and identification benchmarks of the RoboCup@Home research community.FONDECYT-Chile 109025

In this work it is described a framework for classifying face images using Adaboost and domain-partitioning based classifiers. The most interesting aspect of this framework is the capability of building classification systems with high accuracy in dynamical environments, which achieve, at the same time, high processing and training speed. We apply this framework to the specific problem of gender classification. We built several gender classification systems under the proposed framework using different features (LBP, wavelets, rectangular, etc.). These systems are analyzed and evaluated using standard face databases (FERET and BioID), and a new gender classification database of real-world images.