基于优化BP神经网络的手写体识别系统

Handwriting Recognition System Based on Optimized BP Neural Network

在日常生活中手写体随处可见,随着科学技术的不断发展进步,利用计算机识别手写字符成为趋势,为后续数据处理提供便利。对采集手写图像的特征提取和对处理后图像的分类识别进行研究,采用Gabor算法处理经过Sobel算子预处理后的图像,利用PCA进行降维,减少计算量。介绍了基本的BP神经网络,并针对其缺点提出弹性BP算法并进行优化,引入全局误差的评价标准,使得权值更新具备了适应学习和弹性两种优势,保证了全局收敛的条件下大大提升了算法的训练速度和精度。并针对具体的手写体识别问题,实现手写体识别系统的设计,对两种算法分别进行比较验证,进行数据统计,改进后的弹性BP算法拥有更优良的性能,训练时间大大缩短,准确率提升明显。

Handwritten characters are everywhere in our daily life.With the continuous development and progress of science and technology,it has become a trend to recognize handwritten characters by computer,which provides convenience for subsequent data processing.In this paper,the feature extraction of captured handwritten images and the classification and recognition of the processed images are studied.Gabor algorithm is used to process the pre-processed images by Sobel operator,and PCA is used to reduce the dimension and reduce the computational complexity.In this paper,the basic BP neural network is introduced,and the elastic BP algorithm is proposed and optimized according to its shortcomings.The evaluation criterion of global error is introduced,which makes the weight updating have the advantages of adaptive learning and elasticity,and greatly improves the training speed and accuracy of the algorithm under the condition of global convergence.Aiming at the specific handwriting recognition problem,the design of handwriting recognition system is realized.The two algorithms are compared and validated,and the data are counted.The improved elastic BP algorithm has better performance,shorter training time and higher accuracy.