Pattern Recognition Technology Research On The Application Of Biological Engineering

This paper uses the pattern recognition technique wavelet analysis methodand the BP neural network technology, applies to the CT image recognition and3D face recognition in the biological engineering. Discusses the patternrecognition, CT image recognition,3D face recognition research presentsituation and the existing problems. Introduces the key basic theory andtechnology, analyzes the brain CT graphics anomaly detection methods.Researches the some aspects of key technologies which involved the3D facerecognition: Correct posture, facial identification and so on. In order to facilitateresearch, designs and implements the relevant prototype system. Specificcontent as follows:1. This paper proposes a Hough transform based on random and BPnetwork of brain anomaly detection algorithm for cerebral infarction CT imagesautomatic reading problems. First through the random Hough transform, usesdifferent shade gray information of brain CT images, tests round or kind ofcircular area, preliminary judgment possible abnormal area, the area centerposition and radius can be calculated by using RHT transform. Then acquiresgray feature information of the probable abnormal spots, using BP neural network trained model analyzes the gray feature information, rules out normalarea, determines brain abnormal spot.2. This paper proposes the nose position detection algorithm BCCNPDbased on the curvature transform. By using wavelet singularity detectionprinciple to determine the three-dimensional space nose position, realizesaccurate positioning nose. Through the BCCNPD algorithm, uses the spacecharacteristics of bridge of the nose, adjusts the face pose. The experimentshows that the nose position detection algorithm BCCNPD can adjust all sorts ofattitude of3D face data to the correct attitude. Uses the face which is adjustedby the nose position detection algorithm BCCNPD, through plane parallel toxoy plane do the cut surface, Can directly eliminate ears and neck redundantdata and so on, accurate extract3D face outline, the test effect is good.3. This paper proposes based on wavelet analysis3d face feature extractionalgorithm WAFFE, Using Daubec-hies wavelet extract the tangent curvaturechange point of the face facets, realizes the face feature points automaticorientation and extraction; Using WAFFE algorithm can realize the large-scale3d face data storage and retrieval; Based on WAFFE algorithm gets the rapidfeatures comparison algorithm FFMA, it can simplify face feature comparisonprocess, the test effect is good.