A New Approach based on Image Processing Techniques for Automatic Attendance System with Audio Output

The traditional way of taking attendance in a classroom becomes a difficult task and also it is time consuming. In order to overcome such difficulties a new approach is proposed for automatic attendance system. This automatic attendance system detects the student face automatically in the class room and mark the attendance by recognizing the face of student. The proposed approach consists of 3 phases. In phase 1, face is detected using Viola-Jones algorithm. In phase 2, features are extracted by applying principal component analysis. Phase 3, is face recognition process which recognizes the face by finding Euclidean distance between extracted features of detected face in input image with the features of face images in database by using K-Nearest Neighbor. The attendance of student is marked based on the output of phase 3. The absentees list is said loudly through voice conversion system. Experiments are conducted for different input images. The results obtained shows that face detection accuracy is 89 % and for face recognition accuracy is 72%.