Generating statistically significant datasets for face matching system evaluation is a laborious and expensive process. Capturing variables such as atmospheric turbulence and other weather conditions especially with respect to face recognition at a distance exacerbate the problem further. It is even more difficult to work on system issues for long-range systems that impact the collection phase such as automated control loops for gain, focus or zoom, as they directly impact the collected data. And since system performance is confounded with variations in subject selection, pose, lighting, expression, etc., formal evaluation of second order effects are difficult without extremely large collections.This paper describes a taxonomy of face-models for controlled experimentation that overcome these challenges. We show that a gap has existed in experimental design and how a range of model-based approaches can partially fill that gap. Methods for generating 3D models that can be easily manipulated to create variations in pose are presented. Additionally described are techniques for validating and capturing model-based data for use in developing and testing outdoor long-range face matching systems.
No abstract
Abstract-Challenges for face recognition still exist in factors such as pose, blur and distance. Many current datasets containing mostly frontal images are regarded as being too easy. With obviously unsolved problems researchers are in need of datasets that test these remaining challenges. There are quite a few datasets in existence to study pose. Datasets to study blur and distance are almost non-existent. Datasets allowing for the study of these variables would prove to be useful to researchers in biometric surveillance applications. However, until now there has been no effective way to create datasets that encompass these three variables in a controlled fashion.Toolsets exist for testing algorithms, but not systems. Designing and creating toolsets to produce a well controlled dataset or to test the full end-to-end recognition system is not trivial. While the use of real subjects may produce the most realistic dataset, it is not always a practical solution and it limits repeatability making the comparison of systems impractical. This paper attempts to address the dataset issue in two ways. The foremost is to introduce a new toolset that allows for the manipulation and capture of synthetic data. With this toolset researchers can not only generate their own datasets, they can do so in real environments to better approximate operational scenarios. Secondly, we provide challenge datasets generated from our validated framework as a first set of data for other researchers. These datasets allow for the study of blur, pose and distance. Overall, this work provides researchers with a new ability to evaluate entire face recognition systems from image acquisition to recognition scores.
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.