作为深度前馈人工神经网络的一种,卷积神经网络在图像识别领域得到了成功应用.其中,最经典的卷积神经网络模型就是LeNet-5模型.在MNIST字符库上运用该模型,通过优化卷积层的样本训练方式,即将原来以每批固定输入样本数量、固定迭代次数的训练方式,优化为以每批不同输入样本数量、不同迭代次数的混合训练样本方式.优化后的训练方式能够减少预处理工作量,加快识别速度.实验结果表明:在保证样本训练时间相等的前提下,优化后的混合样本输入方式可以得到更高的识别率.
As a kind of depth feedforward artificial neural network,convolutional neural network has been successfully ap?plied in the field of image recognition. Among them,the LeNet-5 model is the most classic convolutional neural network model. This model is used on the MNIST character library and the sample training method of convolution layer is optimized. That is to say, the training method that uses the number of fixed input samples per batch and the number of fixed iterations is optimized to be a mixed training sample mode with different numbers of input samples per batch and different iterations. The optimized training meth?od can reduce the pre-processing workload and speed up the recognition speed. The experimental results show that the optimized mixed sample input method can get a higher recognition rate under the premise of equal sample training time.