Biometric Time Attendance: Use of Bayesian Approach

Large corporate houses, government offices, schools, hospitals, and other places do not mark time attendance for the employees manually. Days of keeping pen and paper based attendance registers are gone. Everything is getting digital today. So is this! Biometry is used in accomplishing this task. Fingerprint examination, voice recognition, iris scan, palm reading, etc. are a few of the tactics used for identity verification. The most popular and the most progressing out of them all is facial detection. Face recognition time attendance has already established a strong hold in the biometric time attendance market. Its use is expected to grow in the near future. rnFace recognition time attendance can employ various algorithms. The algorithms for identification can be image based or video based. Generally, image based algorithm is used for a device recording time and attendance details of the workforce. Principle Component Analysis (PCA), Linear Discriminatory Analysis (LDA), Elastic Bunch Graph Matching, etc. are few out of the many. Bayesian framework is in discussion nowadays! It is a subspace based face recognition approach. PCA, LDA, and Bayesian framework are the three major approaches in this category. The use of Bayesian approach is particularly growing in the statistical community. Special attention is given to its use in biometric equipments. rnApplication of Bayesian approach has become talk of the town. Canadian researchers are making use of this algorithm in face recognition time attendance to get more accuracy. Certain external factors like illumination changes, facial expressions, and face poses may influence detection. Thermal characteristics, however, remain unchanged. This algorithm makes use of this trait. Unlike expectation maximization algorithm, it does not try to take into account the single best model. It considers the average result computed over several models. Researches make use of this approach in combination with other tactics to acquire maximum output out of a biometric time attendance. rnEvery coin has two sides to it. Face recognition time attendance using Bayesian approach has its own disadvantages. It is difficult to implement. Bayesian engineers have to rack their brains hard behind a device using this algorithm. Results are not objective. They are not independent. Prior information may not be accurate and may give misleading conclusions. There is no one “correct” way of inputting prior information and different approaches can give different results. Customers may not accept validity of prior data or engineering judgments. rnThis is one of the approaches used in face recognition time attendance. It is gaining quick popularity in this market. Researchers are experimenting in a variety of ways with this technology. Various algorithms are used in combination with each other to achieve best results. The biggest advantage of doing these efforts is a scope for high accuracy. Probability for false accept rate and false reject rate is reduced in a biometric time attendance. One may easily rely on this instrument. All the algorithms are designed to provide enhanced accuracy. Bayesian approach is no exception! Though it is not as renowned and accepted as the PCA and LDA, it is making a mark for itself at a fast pace.