Research Of Palmprint Segmentation Algorithm Based On Palmprint Diagnosis

In recent years, the biological recognition technology which based on palmprint feature ofattaches great importance to the researchers at home and abroad. Because biological recognitiontechnology based on palmprint features and Palmprint Diagnosis technology are both based oncharacteristics of hand analysis, so we can draw on the relevant biometric technology forautomatic Palmprint Diagnosis research.In Palmprint Diagnosis,one of the most important diagnostic features are different pathologicstructures formed by palm lines in different field.These pathological pattern only appear in a fixedarea have pathological significance, so accurately on the palm image segmentation andsegmentation is a key prerequisite for the realization of the palm pathological patternextraction.This dissertation carried in-depth research on hand image collection, pre-processing,and positioning segmentationof the Palmprint Diagnosis automatic diagnostic system,achievedcertain results,the main work and contributions are summarized as follows:1. For the extraction of a binary image contour portion, on the basis of the analyticalmethods of the8-neighborhood improvements proposed based on the binarized image of thecontour extraction method of the palm of the four templates.Experiments show that the method isnot only simple, fast, effective, but also suitable for all binarization image contour extraction.2.Since the acquisition palm image does not limit the palm of the placement and the fingeropening degree, thus increasing the difficulty of the palmprint image positioning and region ofinterest extraction.In view of the existing problems of palmprint image location and segmentationalgorithm,this dissertationproposed a method to establish the Cartesian coordinates positioningusing connection point in the palm of your hand and wrist.The method can not only quickly tocomplete the establishment of the coordinate system, but also can remove the wrist part, whichcleared the way for subsequent processing.3.In the part of the palm region of interest extraction, this paper proposes a parallel to thecutting line with the palm of your hand corner of the intersection of the contour detectionmethod.This method is not only able to quickly extract the corner when the fingers open, but alsois a better solution to the problem of the fingers together when some commonly used methods cannot be completed, and ultimately the use of these corner to the completion of the removal of part of a finger, successfully extracted Palmprint Diagnosis full palm image information.4.After analysis the characteristic and classification standard of the palm Medical EightDiagrams the paper defines some key points related lines,and use them draw the palm MedicalEight Diagrams on full palm image,then marked the nine regional connectivity of the palmMedical Eight Diagrams.Finally,put forward a method of through the centroid coordinates tolocate the position and the connected region segmentation.Experiments show that the methodapplied with good accuracy and robustness.