| 摘要 |
PROBLEM TO BE SOLVED: To provide a method and an apparatus for recognizing shapes from three-dimensional data. SOLUTION: Feature data is extracted from three-dimensional data using three-dimensional higher-order local autocorrelation feature extraction. The feature data extracted is transformed using a statistical technique such as multivariate analysis to create new feature data. The new feature data is compared with registered data and judged. Feature values extracted by the three-dimensional higher-order local autocorrelation feature extraction are values with invariable positions on a target solid that do not depend on location and time. If there are a plurality of targets in the solid, an overall feature value is the sum of individual feature values and is thus easy to handle for subsequent recognition. Furthermore, the amount of calculations required to extract the features is so small that real-time processing is possible. COPYRIGHT: (C)2008,JPO&INPIT
|
