基于改进型稀疏自动编码器的图像识别OA
Image Recognition Based on Improved Sparse Auto-encoder
传统的稀疏自动编码器不具备平移不变性, 同时对非高斯噪声较为敏感. 为增加网络平移不变的特性,借鉴卷积神经网络的相关理论, 通过对原始的像素块进行卷积运算以达到上述目的; 而为了提高对非高斯噪声的鲁棒性, 自动编码器的代价函数由均方误差改为了最大相关熵准则. 通过在MNIST和CIFAR-10 数据集上进行试验, 结果证明, 改进后的方法较传统的自动编码器具有更好地识别效果, 识别率提高了2%~6%.
The traditional sparse auto-encoder lacks invariant translation and is sensitive to non-Gauss noise .A method convolving the original pixel block is proposed to increase the network invariance with the mean square error (MSE) replaced by the maximum correntropy criterion (MCC) in cost function to improve the anti-noise ability. The proposed method is evaluated using the MINIST and CIFAR-10 datasets .Experimental results show that the pro-posed approach impro…查看全部>>
尹征;唐春晖;张轩雄
上海理工大学光电信息与计算机工程学院,上海 200093上海理工大学光电信息与计算机工程学院,上海 200093上海理工大学光电信息与计算机工程学院,上海 200093
信息技术与安全科学
深度学习自动编码器卷积神经网络最大相关熵
deep learningauto-encoderconvolutional neural networkmaximum correntropy criterion
《电子科技》 2016 (1)
124-127,4
评论