About the Customer
The UAE based client is one of the largest household manufacturers serving a large customer base across the globe. They have more than 10k employees working in various manufacturing units, distribution centers, warehouses, etc.
Due to such a massive employee base, the client had concerns about tracking the attendance of the employees.
The challenges they faced were:
- Inability to monitor in real-time if employees are present
- The huge administrative effort to process paper attendance
- Late payroll
- Unable to prevent time theft or buddy punching
- Chances of manual errors due to heavy paperwork
Recognizing the challenges of the client, SPEC INDIA came up with the AI-powered face recognition solution called SAFR (Secure Accurate Facial Recognition). It uses computer vision and machine learning algorithms to mark the attendance of the employees.
The goal was to introduce a high-tech solution that would eliminate the challenges and provide real-time attendance views.
Face recognition is achieved using machine learning and the basic pipeline used for it is as follows:
- Finds face in an image.
- Analyses facial features.
- Compares against known faces and makes a prediction
Solution Works in the following manner
- Cameras installed at the door detect the face
- SAFR will compare the detected face with the images in the database and make the prediction
- Will mark the attendance and open the door only if the detected face is matched with the image in the database
- Tracking of total time spent in the office
- Data for the payroll system
The technology works by scanning a person’s face “in the wild”, means the system will detect the faces even when it is in motion, under poor lighting or at different angles.
Through this system, the admin will be able to perform below task
- Calculate Salary
- Attendance Report
Tools and Technologies
- Easy tracking of employees
- Employees don’t need to carry extra things for attendance or door opening
- Quick and efficient verification of an employee
- Use of infrared cameras significantly boosted the level of accuracy
- More reliable and harder to trick
- Easy to integrate with existing security infrastructures
- Increased productivity of administrative department as they got rid of time-consuming manual attendance
- Inherently vulnerable to proxies and manual errors