Content of this page

• Introduction
• Why FaceCell?
• Algorithm
• Specifications
• Trial version
• Related products

Face Recognition Technology for Embedded Systems

Introduction

Like fingerprint recognition, human face-based biometrical identification is becoming increasingly popular. Facial recognition is used in systems that control access to physical locations, computer/network resources, bank accounts, or register employee attendance time in enterprises. Many of these applications can run on a PC. However some applications require that the system be implemented on low cost, compact and/or mobile embedded devices such as cell phones, handheld PCs, door or gate locks, etc.

In comparison with fingerprints, facial recognition in embedded devices can be even more practical in many situations and more comfortable for the user because no physical contact with the device is required. PDAs, smart phones and other compact devices with integrated cameras and the ability to add custom software are available in the market, enabling the implementation of embedded facial recognition technology without additional hardware development.

The facial recognition algorithm for embedded systems requires some specific features in comparison with PC-based face identification. Embedded or handheld devices usually have weaker processors than personal computers. The PC-based face image detection and template extraction software is computationally expensive; therefore substantial algorithm modification is required to achieve acceptable template extraction time on the embedded device.

Having the ability to create a mixed embedded/PC and/or multi-biometric face/fingerprint identification system by integrating several technologies is an important advantage. Using a combination of biometric technologies allows implementation of systems with higher levels of security and reliability, as well as achieving higher matching speeds even when using very large databases.

This page contains information about the FaceCell embedded facial recognition technology, developed on the VeriLook basis, but having about 3 times faster image processing and feature extraction algorithm.

Why FaceCell?

FaceCell algorithm is designed for embedded biometric systems developers. The algorithm has certain capabilities:

Click to zoom

Reliability. The FaceCell technology is intended for hardware with lower computational capabilities than PCs. Compared to the PC-based VeriLook 3.0 algorithm, the FaceCell algorithm has a higher, but acceptable False Rejection Rate. The graphical chart compares FaceCell ROC with VeriLook 3.0 ROC using face images from XM2VTSDB database. Identification ability. FaceCell is designed not only for verification (1:1 matching), but also for identification (1:N matching). The algorithm is able to match up to 3,000 faces per second. Easy integration. FaceCell can be used in a wide range of applications and can be easily integrated into handheld or embedded devices with built-in video cameras, such as PDAs and smart phones, without having to develop any special hardware. Portability. FaceCell Embedded Development Kit is designed for easy implementation into very various and specific applications. The algorithm's ANSI C source code can be ported to various platforms and hardware. Embedded and PC-based multi-biometric capable technologies from the same vendor. Combined with our other technologies, FaceCell could be used in developing these advanced systems: Multi-biometric embedded systems, using FaceCell EDK together with FingerCell EDK. Mixed embedded/PC systems, using FaceCell EDK together with VeriLook Standard SDK. Complex multi-biometric embedded/PC systems, using a combination of FaceCell EDK, FingerCell EDK, VeriLook SDK and VeriFinger SDK.

Algorithm

The FaceCell algorithm is similar to the VeriLook algorithm and includes these features:

• Fast and accurate face localization for reliable detection of multiple faces in the images.
• Simultaneous multiple face processing and identification in a single frame. All faces in the current frame are detected in about 1 second^* and then each face template is extracted in about 1 second^*.
• The FaceCell face template matching algorithm compares 3,000 faces per second^*.
• Applications implemented using FaceCell EDK can handle large face databases, as one facial feature template is only 2.3 Kbytes.
• Features generalization mode generates the collection of the generalized face features from several images of the same subject. Then each face image is processed, features are extracted, and the collections of features are analyzed and combined into a single generalized features collection which is written to the database. This way, the enrolled feature template is more reliable and the face recognition quality increases considerably.

^* All performance evaluations were performed using a HP iPAQ Pocket PC with XScale PXA270 processor running at 416 MHz

Specifications

All performance evaluations were performed using a HP iPAQ Pocket PC with XScale PXA270 processor running at 416 MHz

FaceCell trial version

Neurotechnologija offers FaceCell EDK on a 30 day trial. This kit allows developers to explore the technology and to try it in real environments and real applications.
FaceCell EDK trial is available for downloading.

Related products

• FaceCell Library EDK
• FaceCell source code EDK