In automobile industry, the research and development of ITS (Intelligent Transport Systems) is carried out leveraging technology to maximize safety as the first priority. As the fatigue and stress are easy to induce the driver to an accident, several ITS use the human facial expression to guess the driver's psychological condition. However, the detection accuracy of the facial characteristic points decreases greatly when there is a shadow on the face. The method of correcting only the shadow is not devised. This study automatically corrects the brightness of the shadow area, and improves the detection accuracy of the characteristic points. The proposed method and the method using histogram equalization were compared. A significant improvement is obtained by proposed method when the area of the shadow in the face image is 40% or less.
|Number of pages||8|
|Journal||Journal of the Institute of Image Electronics Engineers of Japan|
|Publication status||Published - 1 Jan 2012|
- Correction of lightness
- Facial expression
- Image processing
- Light-source direction