The acceptance of service robots comes along with the ability to adapt to user specific preferences. This requires that a robot can determine the identity of the user. As for humans, robust user recognition is based on the identification of the face. However, despite the plethora of published work on face recognition that is robust against real world noise such as illumination, head alignment or facial expressions there is no robust off-the-shelf non-commercial software available to be used in typical robotics applications. Hence, this paper introduces a ready-to-use open-source ROS package providing a face detection and identification system that is comprising novel and state-of-the-art solutions to various aspects of face recognition while utilizing modern RGB-D sensors. This work demonstrates a solution for face recognition in robotic settings that is robust against varying illumination, gaze directions of the head, and facial expressions while operating with online performance. The paper provides a thorough evaluation of the face recognition system based on standard database tests and on real world scenarios regarding these criteria. The work described in this project was partially funded by the European project ACCOMPANY (Acceptable robotics COMPanions for AgeiNg Years). Grant agreement no.: 28762
ABSTRACT:The paper presents the implementation of a dense multi-view stereo matching pipeline for the evaluation of image sequences from a camera-based mobile mapping system. For this purpose the software system SURE is taken as a basis. Originally this system was developed to provide 3D point clouds or DEM from standard airborne and terrestrial image blocks. Since mobile mapping scenarios typically include stereo configurations with camera motion predominantly in viewing direction, processing steps like image rectification and structure computation of the existing processing pipeline had to be adapted. The presented investigations are based on imagery captured by the mobile mapping system of the Institute of Geomatics Engineering in the city center of Basel, Switzerland. For evaluation, reference point clouds from terrestrial laser scanning are used. Our first results already demonstrate a considerable increase in reliability and completeness of both depth maps and point clouds as result of the matching process.
We describe a novel framework that combines an overhead camera and a robot RGB-D sensor for real-time people finding. Finding people is one of the most fundamental tasks in robot home care scenarios and it consists of many components, e.g. people detection, people tracking, face recognition, robot navigation. Researchers have extensively worked on these components, but as isolated tasks. Surprisingly, little attention has been paid on bridging these components as an entire system. In this paper, we integrate the separated modules seamlessly, and evaluate the entire system in a robot-care scenario. The results show largely improved efficiency when the robot system is aided by the localization system of the overhead cameras
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.