The B-spline curve and surface provide an accurate tool to record object shape. We present a biometric identification system through hand geometry measurements by using B-spline curves. We use 4 B-spline curves to fit with fingers (except thumb) from a single hand image for a single person. Then we store these 4 curves as well as other geometry measurements of the hand as the "signature" of that person into the database. By computing the differences between the curves from database hand images and the curves from the query hand image using the point projection method, we are able to verify/identify the person by locating the closest database hand image to the query hand image.
|Title of host publication||Proceedings - International Conference on Pattern Recognition|
|Publisher||The Institute of Electrical and Electronics Engineers (IEEE)|
|Number of pages||4|
|Publication status||Published - 2004|