Sketch-based Facial Expression Recognition for Human Figure Drawing Psychological Test
| Autor: | Tasneem Aslam, Imran Siddiqi, Momina Moetesum, Hassan Saeed, Uzma Masroor |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Local binary patterns
business.industry Computer science Feature extraction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 020206 networking & telecommunications 02 engineering and technology computer.software_genre Facial recognition system Expression (mathematics) Sketch Support vector machine Histogram of oriented gradients 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Psychological testing Artificial intelligence business computer Natural language processing |
| Zdroj: | FIT |
| DOI: | 10.1109/fit.2017.00053 |
| Popis: | Drawing tests have been long used by practitioners for early screening of a number of psychological and neurological impairments. These brain functioning tests are used by psychologists to understand feelings, personality and reactions of individuals to different circumstances. Among these, Human Figure Drawing Test (HFDT) is a popular instrument for the assessment of cognitive functioning of individuals. While the HFDT has various dimensions, the focus of this study lies on the face of the drawn figure. A computerized system that analyzes the hand-drawn facial images to extract the expressions from the image is proposed. Sketch of human face is drawn by the subject and then fed to the system, the image is then binarized and segmented into different facial components. Features (based on local binary patterns, gray level co-occurrence matrices and histogram of oriented gradients) computed from the facial components are used to train an SVM classifier to learn to distinguish between four expression classes, ‘happy’, ‘sad’, ‘angry’ and ‘neutral’. The system evaluated on a custom developed database of sketches realized promising results. The developed system could serve as a useful module toward development of a complete automated system to score human figure drawing test. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|