Download Full Text (441 KB)
Devanshi Malaviya, CIS435: Python Programming
Faculty Mentor(s): Professor Sarbani Banerjee, Computer Information Systems
Face recognition in real time is a popular subject of research and a rapidly growing challenge. Emotion recognition by focusing on various parts of the face and speech tones is also an exciting field. In this project, my focus is on face recognition and smile detection. I am planning to build a camera based on a real-time face recognition system that can detect if a person is smiling or not and display a message to that effect. The Python packages used in developing this project are HaarCascade, OpenCV, LBPH, numpy and PIL. The face recognition is done in 3 steps, starting with face detection, then using the LBPH algorithm for identification and verification of facial image. The LBPH algorithm uses a pixel matrix on a grayscale image to give new binary value to each cell and in the end, it produces a new image which represents better characteristics than the original picture. The HaarCascade algorithm used for face and smile detection uses a set of elementary combinations of dark and bright areas by edge, linear and central features. In the future, I plan to expand this project and implement it in video-calling software applications. It will use automatically generated captions, along with face and emotion recognition, and generate a transcript or a summary of a meeting after it ends.
Malaviya, Devanshi, "Face Recognition and Smile Detection" (2021). Computer Information Systems and Engineering Technology. 14.