FIRST: Face Identity Recognition in SmarT Bank

2016 ◽  
Vol 10 (04) ◽  
pp. 569-591 ◽  
Author(s):  
Qianmu Li ◽  
Tao Li ◽  
Bin Xia ◽  
Ming Ni ◽  
Xiaoqian Liu ◽  
...  
Keyword(s):  
Face Recognition ◽  
Information Acquisition ◽  
Rapid Development ◽  
Integrated System ◽  
Front Face ◽  
Identity Verification ◽  
Training Time ◽  
Identity Recognition ◽  
Face Identity ◽  
Self Service

The rapid development of information era has influenced to realize the notion of Smart Bank via approaches like paperless services and interactive self-service systems. Since the traditional methods of identity verification are insecure and cumbersome for supporting these services, Smart Bank has been questioned. To overcome the limitations of current identity verification, it is imperative to explore an effective recognition strategy considering the trade-off between security and customer experience, which can conveniently collect identity information and accurately distinguish people. However, few research and existing systems have been reported for an integrated solution of identity verification satisfying convenient and secure banking environment. In this paper, we propose, implement, and deploy an integrated system named FIRST (Face Identity Recognition in SmarT Bank), which is a customized platform for the identity verification via face recognition in Smart Bank. FIRST uses Gabor surface feature and Fisherfaces, which can provide accurate face recognition within acceptable training time. For information acquisition, our system employs the patch-based face quality assessment, which efficiently extracts valuable faces (i.e. front face) from video streams. Furthermore, FIRST provides a distributed environment to effectively manage recognition tasks and massive data. Since March 2015, FIRST has been successfully deployed on 1800 self-service terminals in Jiangsu Province by ABC (Agriculture Bank of China), and is under deployment by State Grid in China.

scholarly journals Computational insights into human perceptual expertise for familiar and unfamiliar face recognition

2019 ◽  
Author(s):  
Nicholas Blauch ◽  
Marlene Behrmann ◽  
David C. Plaut
Keyword(s):  
Face Recognition ◽  
Theoretical Work ◽  
Fine Tuning ◽  
Identity Verification ◽  
High Performing ◽  
Face Identity ◽  
Unfamiliar Face ◽  
Perceptual Representations ◽  
Expert Networks ◽  
Identity Perception

Humans are generally thought to be experts at face recognition, and yet identity perception for unfamiliar faces is surprisingly poor compared to that for familiar faces. Prior theoretical work has argued that unfamiliar face identity perception suffers because the majority of identity-invariant visual variability is idiosyncratic to each identity, and thus, each face identity must be learned essentially from scratch. Using a high-performing deep convolutional neural network, we evaluate this claim by examining the effects of visual experience in untrained, object-expert and face-expert networks. We found that only face training led to substantial generalization in an identity verification task of novel unfamiliar identities. Moreover, generalization increased with the number of previously learned identities, highlighting the generality of identity-invariant information in face images. To better understand how familiarity builds upon generic face representations, we simulated familiarization with face identities by fine-tuning the network on images of the previously unfamiliar identities. Familiarization produced a sharp boost in verification, but only approached ceiling performance in the networks that were highly trained on faces. Moreover, in these face-expert networks, the sharp familiarity benefit was seen only at the identity-based output layer, and did not depend on changes to perceptual representations; rather, familiarity effects required learning only at the level of identity readout from a fixed expert representation. Our results thus reconcile the existence of a large familiar face advantage with claims that both familiar and unfamiliar face identity processing depend on shared expert perceptual representations.

scholarly journals Individual Differences in Serial Dependence of Facial Identity are Associated with Face Recognition Abilities

2019 ◽  
Vol 9 (1) ◽  
Author(s):  
Kaitlyn Turbett ◽  
Romina Palermo ◽  
Jason Bell ◽  
Jessamy Burton ◽  
Linda Jeffery
Keyword(s):  
Face Recognition ◽  
Individual Differences ◽  
Functional Role ◽  
Perceptual Bias ◽  
Serial Dependence ◽  
Identity Recognition ◽  
Face Identity ◽  
Facial Identity ◽  
Visual Attributes ◽  
Recognition Ability

AbstractSerial dependence is a perceptual bias where current perception is biased towards prior visual input. This bias occurs when perceiving visual attributes, such as facial identity, and has been argued to play an important functional role in vision, stabilising the perception of objects through integration. In face identity recognition, this bias could assist in building stable representations of facial identity. If so, then individual variation in serial dependence could contribute to face recognition ability. To investigate this possibility, we measured both the strength of serial dependence and the range over which individuals showed this bias (the tuning) in 219 adults, using a new measure of serial dependence of facial identity. We found that better face recognition was associated with stronger serial dependence and narrower tuning, that is, showing serial dependence primarily when sequential faces were highly similar. Serial dependence tuning was further found to be a significant predictor of face recognition abilities independently of both object recognition and face identity aftereffects. These findings suggest that the extent to which serial dependence is used selectively for similar faces is important to face recognition. Our results are consistent with the view that serial dependence plays a functional role in face recognition.

scholarly journals Deep-Learning-Based Remaining Useful Life Prediction Based on a Multi-Scale Dilated Convolution Network

Mathematics ◽  
2021 ◽  
Vol 9 (23) ◽  
pp. 3035
Author(s):  
Feiyue Deng ◽  
Yan Bi ◽  
Yongqiang Liu ◽  
Shaopu Yang
Keyword(s):  
Deep Learning ◽  
Rapid Development ◽  
Operational Efficiency ◽  
Remaining Useful Life ◽  
Convolution Kernel ◽  
Training Time ◽  
Multi Scale ◽  
Dilated Convolution ◽  
Convolution Kernels ◽  
Useful Life

Remaining useful life (RUL) prediction of key components is an important influencing factor in making accurate maintenance decisions for mechanical systems. With the rapid development of deep learning (DL) techniques, the research on RUL prediction based on the data-driven model is increasingly widespread. Compared with the conventional convolution neural networks (CNNs), the multi-scale CNNs can extract different-scale feature information, which exhibits a better performance in the RUL prediction. However, the existing multi-scale CNNs employ multiple convolution kernels with different sizes to construct the network framework. There are two main shortcomings of this approach: (1) the convolution operation based on multiple size convolution kernels requires enormous computation and has a low operational efficiency, which severely restricts its application in practical engineering. (2) The convolutional layer with a large size convolution kernel needs a mass of weight parameters, leading to a dramatic increase in the network training time and making it prone to overfitting in the case of small datasets. To address the above issues, a multi-scale dilated convolution network (MsDCN) is proposed for RUL prediction in this article. The MsDCN adopts a new multi-scale dilation convolution fusion unit (MsDCFU), in which the multi-scale network framework is composed of convolution operations with different dilated factors. This effectively expands the range of receptive field (RF) for the convolution kernel without an additional computational burden. Moreover, the MsDCFU employs the depthwise separable convolution (DSC) to further improve the operational efficiency of the prognostics model. Finally, the proposed method was validated with the accelerated degradation test data of rolling element bearings (REBs). The experimental results demonstrate that the proposed MSDCN has a higher RUL prediction accuracy compared to some typical CNNs and better operational efficiency than the existing multi-scale CNNs based on different convolution kernel sizes.

Development of an antibiotic resistance monitoring system in Hungary

2002 ◽  
Vol 50 (2) ◽  
pp. 189-197 ◽  
Author(s):  
Éva J. Kaszanyitzky ◽  
A. Tarpai ◽  
Sz. Jánosi ◽  
Keyword(s):  
Antimicrobial Resistance ◽  
Animal Species ◽  
Rapid Development ◽  
Inhibition Zone ◽  
Integrated System ◽  
Natural Resistance ◽  
Animal Origin ◽  
Diffusion Method ◽  
Cross Resistance ◽  
Central Computer

Because of the rapid development and spread of antimicrobial resistance it is important that a system be established to monitor antimicrobial resistance in pathogenic zoonotic and commensal bacteria of animal origin. Susceptibility testing of bacteria from carcasses and different samples of animal origin has been carried out in veterinary institutes for a long time but by an inconsistent methodology. The disc diffusion method proposed by the National Committee for Clinical Laboratory Standards (NCCLS) was introduced in all institutes in 1997. In order to obtain a coherent view of the antimicrobial resistance of bacteria a computer system was consulted, consisting of a central computer to store all data and some local computers attached to it through the network. At these local measuring stations computers are connected to a video camera, which displays the picture of Petri dishes on the monitor, and inhibition zone diameters of bacteria can be drawn with the mouse by the inspector. The software measures the diameters, evaluates whether or not the bacteria are sensitive, and stores the data. The evaluation is based upon the data of the NCCLS. The central computer can be connected to as many local computers with measuring stations as we wish, so it is suitable for an integrated system for monitoring trends in antimicrobial resistance of bacteria from animals, food and humans, facilitating comparison of the occurrence of resistance for each circumstance in the chain. It depends on the examiners which antibiotics they want to examine. Thirty-two different antibiotic panels were compiled, taking into consideration the active ingredients of medicinal products permitted for veterinary use in Hungary, natural resistance and cross-resistance, the mechanism of resistance and the animal species, i.e. which drugs were recommended for treatment in the given animal species, and the recommendations of the OIE Expert Group on Antimicrobial Resistance. The members of the panels can be changed any time, even during the measuring process. In addition to the inhibition zone diameters of bacteria the database also includes information about bacterial and animal species, the age of animals and the sample or organ where the bacteria are from. Since January 2001 the antibiotic susceptibility of E. coli, Salmonella, Campylobacter and Enterococcus strains isolated from the colons of slaughter cows, pigs and broiler chickens has also been examined. Each of the 19 counties of Hungary submits to the laboratory three tied colon samples from a herd of the above-mentioned animals every month.

scholarly journals Accelerating the Social Media Process: The Impact of Internet Celebrity Word-of-Mouth Communication and Relationship Quality on Consumer Information Sharing

2020 ◽  
Vol 10 (1) ◽  
pp. 201
Author(s):  
Liping Liu ◽  
Chih-Cheng Fang
Keyword(s):  
Social Media ◽  
Information Sharing ◽  
Relationship Quality ◽  
Information Acquisition ◽  
Word Of Mouth ◽  
Rapid Development ◽  
Consumer Information ◽  
Research Fields ◽  
Word Of Mouth Communication ◽  
The Impact

With the rapid development of "Internet plus", the number of Internet users in China has increased rapidly, and the number of active users of social media software ranks first in the world. Large Numbers of network users are also potential consumer groups. Social media influences other consumers through consumer interaction and social interaction, and consumers are transformed into active information acquisition rather than passive information reception. Word of mouth marketing on social media has become one of the hottest research fields. Based on the information adoption model, this study explores the impact of internet celebrity word-of-mouth communication on consumer information sharing from four dimensions: internet celebrity word-of-mouth communication, relationship quality, face consciousness, and consumer information sharing and establishes a research model to provide references and suggestions for subsequent researchers and enterprise management.

Research on Data Mining Algorithm Based on Pattern Recognition

2019 ◽  
Vol 34 (06) ◽  
pp. 2059015 ◽  
Author(s):  
Xuelong Zhang
Keyword(s):  
Data Mining ◽  
Pattern Recognition ◽  
Rapid Development ◽  
Future Trend ◽  
Support Vector ◽  
Massive Data ◽  
Data Mining Algorithm ◽  
Rapid Improvement ◽  
Training Time ◽  
Database Technology

With the advent of the era of big data, people are eager to extract valuable knowledge from the rapidly expanding data, so that they can more effectively use these massive storage data. The traditional data processing technology can only achieve basic functions such as data query and statistics, and cannot achieve the goal of extracting the knowledge existing in the data to predict the future trend. Therefore, along with the rapid development of database technology and the rapid improvement of computer’s computing power, data mining (DM) came into existence. Research on DM algorithms includes knowledge of various fields such as database, statistics, pattern recognition and artificial intelligence. Pattern recognition mainly extracts features of known data samples. The DM algorithm using pattern recognition technology is a better method to obtain effective information from massive data, thus providing decision support, and has a good application prospect. Support vector machine (SVM) is a new pattern recognition algorithm proposed in recent years, which avoids dimension disaster by dimensioning and linearization. Based on this, this paper studies the DM algorithm based on pattern recognition, and proposes a DM algorithm based on SVM. The algorithm divides the vector of the SV set into two different types and iterates through multiple iterations to obtain a classifier that converges to the final result. Finally, through the cross-validation simulation experiment, the results show that the DM algorithm based on pattern recognition can effectively reduce the training time and solve the mining problem of massive data. The results show that the algorithm has certain rationality and feasibility.

Sign in / Sign up

Export Citation Format

Share Document