基于PCNN网络的印章图像处理  

Process of the Seal Image Based on PCNN

作　　者：林毅[1] 傅夏生 

机构地区：[1]四川大学计算机学院,成都610065

出　　处：《现代计算机（中旬刊）》2017年第2期71-76,共6页Modern Computer

摘　　要：当今社会,出现大量的假印章,假章的泛滥导致了严重的问题,因此对印文图像进行精确而高效的识别就显得非常重要。对PCNN模型进行深入的研究,并着重学习实践应用PCNN对印文图像进行处理研究。脉冲耦合神经网络是和生物智能领域的结合,具有生物神经网络独特的高容错性和高适应性,能够保证印文图像在印文残缺,线条不均匀的情况下不会影响印章的识别,同时也能够满足对印文图像识别的实时性和准确性的要求[1]。应用PCNN进行印文提取质量较高,提取速度快。应用PCNN模型对印文图像进行提取,探究应用人工神经网络和传统上提取印文图像红色分量匹配的结果,更好地理解人工神经网络在图像处理上应用的相关技术。With the development of the society, the seal plays a significant role in today＇s social life as a symbol of the credit. In today＇s society,there are many kinds of fake seals. Because the flooding of fake seals will cause serious problems, it is very important to identify the seal image accurately and efficiently. Studies this model and focuses on the application of the PCNN for the process of the seal image. The PCNN is a combination with biological intelligence. This model has unique high fault tolerance and high adaptability. By applying this model the seal image, it can be recognized under complicated conditions, and at the same time, it can satisfy the seal image recognition of realtime and accuracy requirements. Using the PCNN to extract seal images has higher quality and faster speed for extracting. Applies the PCNN model to extract the seal image and compare to the traditional method which extracts the red channel of the seal image. Through the study and research on PCNN model to get a better understanding of artificial neural network technologies of its application in image processing.

关 键 词：脉冲耦合神经网络 印章图像 印文提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]