Comparative Study Of Fingerprint Classification Based On LVQ And Neural Network

Nowadays,with the rapid development and wide application of computer technology,biometric technology has become the most convenient and secure identity technology.The stability and uniqueness of the fingerprint make the fingerprint recognition technology occupy nearly half of the market in the biometric recognition technology market.Widely used in civil life,commercial products,criminal investigation,genetic engineering and other fields.At present,there are still many problems with fingerprint recognition technology,such as the poor quality of fingerprint image after fingerprint image preprocessing,the recognition rate of fingerprint matching is not high,and the operation speed is slow.This paper compares the common fingerprint recognition technology based on the BP neural network model and the newly proposed LVQ neural network model to find their advantages and disadvantages.It mainly researches the following aspects.In fingerprint image pre-processing,the fingerprint image is segmented first.In fingerprint image enhancement,first normalize,then calculate the pattern,frequency,area mask,and finally filter.In processing the binarized image,the corrosive expansion operation is used to repair the binarized image to remove some holes and noises in the image.After the fingerprint image preprocessing is completed,the position distribution information of the singular points is used to perform fingerprint classification,and then feature extraction is performed from the refined binary image.Finally,experimental results of fingerprint recognition using two neural network models are made.It is concluded that the recognition rate of fingerprint classification recognition effect of LVQ neural network is higher than that of BP neural network.Although the training time is longer,the stability of the system itself is more reliable.