Design and Implementation of Smart Record System Based on Face Detection and Recognition Using Convolutional Neural Networks Method

The use of a manual attendance system has many disadvantages. These drawbacks include the login process, long response time, and lack of effectiveness. The face recognition-based Smart Record System is an effective step for its replacement. This research uses a quantitative approach method. The quantitative approach was chosen because it collects numerical data that can be statistically analyzed to determine the changes and effects resulting from the implementation of the Smart Record System. The results of this study show that the average respondent time is 0s 79ms or 7.9% per second. In testing objects that block the face, the face is not detected if it is 50% blocked and has restrictions on facial positions such as looking to the right, left, and looking up. In the detection distance test, the safe distance is around 30-90cm. In measuring light conditions, the system functions optimally in bright conditions with 120-125 lumen lighting and in dark conditions with lighting around 15-40 lumen. The conclusion of this research is that the application of the Smart Record System is proven to increase efficiency and simplify the attendance process.