Within the last several years, human mood recognition based on facial recognition has been very active research in computer vision. Mood recognition is used for many areas including psychiatrists and education as well. Mood can be recognized through the points of the face, including some measurement such as size, texture, and skin- color changes. This study applied the mood recognition in customer management service as a feedback to the service. Nowadays, customer service usually conduct a manual survey such as feedback form, testimonial and questionnaires to measure the customer satisfaction. Manual customer satisfaction surveys are very subjective and the customer’s response may be less accurate since customer’s behavior is unpredictable. Hence, the objective of this study is to develop a mood recognition prototype as customer satisfaction feedback. This prototype is using Fuzzy Inference System (FIS) as the engine. This study explores the recognition of domain-specific mood using a Fuzzy Inference System (FIS) to detect three categories of mood; negative and positive and neutral based customer’s mouth shape. The finding shows the accuracy is 78% matched from the testing images. For future work, this study will focus on adding more features and face points and improving the rules and also combining other classifiers to the Fuzzy Inference System (FIS) for better performance.
STA005 - MOOD RECOGNITION BASED ON MOUTH-IMAGES USING FUZZY INFERENCE SYSTEM AS CUSTOMER FEEDBACK
Within the last several years, human mood recognition based on facial recognition has been very active research in computer vision. Mood recognition is used for many areas including psychiatrists and education as well. Mood can be recognized through the points of the face, including some measurement such as size, texture, and skin- color changes. This study applied the mood recognition in customer management service as a feedback to the service. Nowadays, customer service usually conduct a manual survey such as feedback form, testimonial and questionnaires to measure the customer satisfaction. Manual customer satisfaction surveys are very subjective and the customer’s response may be less accurate since customer’s behavior is unpredictable. Hence, the objective of this study is to develop a mood recognition prototype as customer satisfaction feedback. This prototype is using Fuzzy Inference System (FIS) as the engine. This study explores the recognition of domain-specific mood using a Fuzzy Inference System (FIS) to detect three categories of mood; negative and positive and neutral based customer’s mouth shape. The finding shows the accuracy is 78% matched from the testing images. For future work, this study will focus on adding more features and face points and improving the rules and also combining other classifiers to the Fuzzy Inference System (FIS) for better performance.
0 Comments