基于改进LeNet-5卷积神经网络的发票识别研究

Research on Invoice Recognition Based on Improved LeNet-5 Convolutional Neural Network

增值税发票信息在金融领域至关重要,自动识别发票信息可以节约人工成本和时间成本。因此,文中提出设计一种基于改进LeNet-5卷积神经网络的发票自动识别方法,将改进LeNet-5卷积神经网络算法应用于发票识别后,经验证表明,应用改进LeNet-5卷积神经网络算法的发票信息识别率得到了提高。

VAT invoice information is very important in the financial field.Automatic identification of invoice information can save labor cost and time cost.Therefore,this paper proposes to design an automatic invoice recognition method based on the improved lenet-5 convolutional neural network.After the improved lenet-5 convolutional neural network algorithm is applied to the invoice recognition,the verification shows that the invoice information recognition rate is improved by using the improved lenet-5 convolutional neural network algorithm.