摘要 |
PROBLEM TO BE SOLVED: To obtain a facial image processor which can specify the eye area of a facial image at an improved processing speed and can discriminate where eyes are opened or shut. SOLUTION: The facial image processor extracts a nostril area from a facial image inputted from a CCD camera 2 by spirally retrieving the facial image by means of a feature-of-multilevel image extracting means 9 using an extracting filter and stores the nostril area in a binarized image memory 10. Then, after specifying an eye retrieving extent on the basis of the nostril position which is the middle point of the centroid position of nostrils in the binarized nostril area stored in the memory 10, the processor extracts an eye area from the facial image by using the extracting filter and stores the eye area in the memory 10. Thereafter, the processor discriminates where eyes are opened or shut from the binarized eye area by means of an opened/shut state discriminating means 11.
|