Human face identification has been considered an exciting exploration area for a decade. It plays a major role in biometric authentication in several operations including attendance management and access control systems. The attendance systems are important to all associations due to their complexity and time-consuming for maintaining regular attendance log records. There are numerous automated human identification ways similar to biometrics, RFID, eye surveillance, and voice recognition. Face is one of the most immensely used biometrics for human authentication. This paper presents a multimodal face recognition system based on deep learning convolutional neural networks and Haar features.  The transfer learning is used with three pre-trained convolutional neural networks and training on the face image dataset. A real-time class attendance system will be a helpful tool for monitoring attendance. This system operates automatically, using a camera to detect human faces as per the images in the face image dataset. An interface has been developed that will be useful for the school administrator to keep track of students’ activities. Machine learning can help the system to learn and make decisions without human intervention.