HUMAN MOOD DETECTOR BY USING DIGITAL IMAGE PROCESSING

• Main Author: ALI, NOORAZLIN
• Format: Final Year Project
• Language: English
• Published: Universiti Teknologi Petronas 2006
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://utpedia.utp.edu.my/6982/1/2006%20-%20HUMAN%20MOOD%20DETECTOR%20BY%20USING%20DIGITAL%20IMAGE%20PROCESSING.pdf; http://utpedia.utp.edu.my/6982/

The human face provides the most salient clue to person's emotional state. The face expressions are recognized -usingfeataies sueii as eye, eyebrow, mouth or nose. These features play a vital role in human mood recognition. This project, which is named as 'Human Mood Detector', will be recognizing a set of emotions, which consists of angry, sad and happy, requiring vast of knowledge in Digital Image Processing. These projects ate divided into two categories, software development in the first semester, and implementation in the next semester. The identification of& person's mood is based on input from digital camera. Then, the input image will be undergo digitization, noise removal, resizing, format transformation, contrast enhancement, segmentation and feature extraction. Veatuie extraction is a crucial stage in this system; since at this stage; the important feature information from the image will be extracted, which will lead to the desired result A few distinct criteria that will distinguish the emotion are recognized, using segmented images of mouth, forehead or eye shape. These distinct criteria will he used to develop an algorithm that differentiates all the three emotions described before. To classify a person's mood, a similarity matching procedure will be performed from the input image whereby the input will be matched with the set of criteriathat has been classified in the program. Thirty images nmU be Tased in the second semester ^witfe different set of mood for testing. If, the eKtracted information falls within the classified features of a certain mood, the mood of the person will be identified. Identification result will be presented at the end withthe probability. Thetheory, concept, proposed methodology as well as softwaredevelopment (coding) arerepresented in this report.