Face Recognition Using Fisherface and Support Vector Machine Method

COVID-19 is declared as a pandemic by WHO and until now COVID-19 pandemic remains a problem in 2021. Many efforts have been made to reduce the spreading virus, one way to reduce its spread is by wearing a mask but most people often ignore it. Monitoring large groups of people becomes difficult by the government or the authorities. Face recognition, a biometric technology, is based on the identification of a face features of a person. This paper describes a face recognition using Fisherface and Support Vector Machine method to classify face mask dataset. Face recognition using Fisherface method is based on Principal Component Analysis (PCA) and Fisher's Linear Discriminant (FLD) method or also known as Linear Discriminant Analysis (LDA). The algorithm used in the process for feature extraction is Fisherface algorithm while classification using Support Vector Machine method. The results show that for face recognition on face mask dataset using cross validation with 10 fold, the average percentage accuracy is 99.76%. © 2021 IEEE.

Network Delay Modelling and Optimization of Internet-Based Distributed Test Platform for Fuel Cell Electric Vehicle Powertrain System

The accelerated global progress in the research and development of automobile products, and the use of new technologies, such as the Internet, cloud computing and big data, to coordinate development platforms in different regions and fields, can reduce the duration and cost of development and testing. Specifically, in the context of the current coronavirus disease (COVID-19) pandemic, which has caused great obstacles to normal logistics and transportation, personnel exchanges and information communication, platforms that can support global operation are significant for product testing and validation, because they eliminate the need for the transportation of personnel and equipment. Therefore, the establishment of a distributed test and validation platform for automotive powertrain systems, which can integrate software and hardware testing, is important in terms of both scientific research and industrialization. The main technical difficulties associated with such test and validation platforms include data transmission and the control of the transmission effect. A distributed test and validation platform for a fuel cell electric vehicle powertrain system is proposed herein. The two-time-scale Markov chain is used to simulate the delay between two places (China and Germany), and the least-squares support vector machine (LSSVM) method is used to optimize the transmission effect. The results show that the two-time-scale Markov chain model can effectively simulate the delay between two nations, and that its probability distribution is close to the measured value. The LSSVM method is effectively optimized for all four indicators (velocity, fuel cell output power, battery output power and electric motor output torque). This method can be effectively used in the remote development test validation of vehicle powertrain system. © 2021 SAE Technical Papers. All rights reserved.