Face Recognition System

Our experience developing a Face Recognition biometric system with minimum human interventions

Customer Profile:

Our client is a major financial service company in Japan. Also, a bank holding company, helps other companies to be made into subsidiaries under the Banking Law. With the business purpose of conducting business that can be done, they carry out business management, necessary to improve the management efficiency of the Group and to realize the combination of features and strengths in business fields and functions.

Problem Statement:

The exponential growth and the increase in their employee count made it difficult for the company to mark the attendance and calculate the payroll manually.

This process of marking attendance, identifying productive working hours, and preparing payrolls based on the same consumes 40% of the teams’ total working time. The chances of a miss-match and the margin of errors can be high in such a scenario.

An online platform that automates employee attendance tracking and can forecast and prepare payrolls with the minimum human intervention will reduce the margin of errors and give highly reliable data..

Project Background:

The client needed a biometric system to avoid any fraudulent activities. We suggested using the face recognition biometric machine to manage employee attendance. The system is capable of calibrating employees’ physiological characteristics and collect their identification like hand shape, fingerprint, iris, and face shape, thus preventing the occurrence of any fraudulent activities. When considering the current fingerprint attendance system as an example, studies have found that the fingerprint attendance system has a margin of error of about 5%. And some time chances are that the fingerprint scanner won’t work efficiently when dealing with large attendance sites and is likely to cause congestion. The chances of employees swiping cards for someone else is too high, and it is difficult to achieve the purpose of real-time attendance. When comparing with the two attendance systems, the face recognition system has higher accuracy and stability, because there are more points for face recognition, which is more accurate than other systems.

Challenges:

  • Intraclass Variation and Interclass Similarity.
  • Recognition under Real-World Settings.
  • Global Representation.

Technology:

Development language: Java

Database Server: AWS RDS – MySQL

File Storage: AWS S3

Content Delivery: AWS CloudFront

Solution:

Innovator proposed a tracking system where employee personal details and their physiological characteristics are mapped onto their respective profiles and equipped to track the activities. The system enables the company to stay connected with their employee and monitor everything about them inside the company, thus bringing a considerable end to the time-theft activities. The Software enables the accurate tracking of employee working time and, the company can pay their employees based on the same.

The system provides accurate and reliable employee attendance. Also helps to resolve buddy punching as it is one of the serious problems faced by many companies with outdated security and tracking systems. Generally, most of the employees will try to save their buddies to get full attendance by punch instead of them. But when it comes to face recognition in Biometric Attendance System, one could not make it possible. Therefore, it is the best technique to follow.

Achievements:

  • Delivered the biometric attendance system from scratch in a short schedule to meet the customer release plan.
  • The system resolved issues like buddy punching and tailgating.

Return on Investment:

Reduced the human intervention in the whole attendance marking and payroll calculation process by 56% thereby saving 40% of the team’s productive time.