Author : Keerti Keshav Kanchi and Vijaya C

The Face recognition is an important and secured way to protect the frauds at everywhere like government agencies which are investing a considerable amount of resources into improving security systems as result of recent terrorist events that dangerously exposed flaws and weaknesses in today’s safety mechanisms. In this paper discrete cosine transform and neural network is used in the development of facial expression recognition system capable of operating in real time. The proposed technique uses video input as input image taken through the webcam in real time on which two dimensional discrete cosine transform (2D-DCT) is performed for image compression and self organizing map (SOM) for recognition purpose. The DCT extracts features from face image; these feature-vectors are constructed by computing DCT coefficients. A self-organizing map(SOM) using an unsupervised learning techniques used to classify DCT-based features vectors into groups to identify if the subject in the input image is present or not present in the image database. Evaluation of the procedure is performed in MATLAB using an image database of 25 people containing 5 subjects and each subject have 5 different facial expressions. After training about 1000 epochs system achieved approximately 89.18% recognition rate.

Keywords:2D-Discrete Cosine Transform (2D-DCT); Facial Expression Recognition; Image Processing Neural Network; Kohonen Self Organizing Map (SOM).

Article published in the Proceedings of National Conference on ‘Women in Science & Engineering’ (NCWSE 2013), SDMCET Dharwad