In human-human communication the face conveys a lot of information. Analyzing faces in human-computer communication is also becoming progressively important in providing security to the users. In existing system the facial expressions are examined and compared with the face images stored in the database. In proposed system TPCF (Tensor Perceptual Color Framework) is used for skin tone color detection and detected color are converted to gray scale value which is used for classifying skin or non-skin. Based on the skin or non-skin detection, the features are extracted and the emotions are calculated which are used for initiating the assigned task. The proposed system identifies the melancholic, smiling, angry emotions, when melancholic emotions are detected alert messages are sent to the attendees and when smiling emotions are detected, the system responds by playing songs etc.