降噪自动编码机在图像识别中的应用  

The Application of Image Retrieval System Based on Noise Reduction Algorithm for Automatic Coding Machine

作　　者：郭海凤[1] 李广水[1] 仇彬任 

机构地区：[1]金陵科技学院信息技术学院,江苏南京211169

出　　处：《金陵科技学院学报》2015年第3期32-35,共4页Journal of Jinling Institute of Technology

基　　金：江苏省高校自然科学研究项目(14KJD520003);金陵科技学院科研基金项目(jit-n-201521)

摘　　要：深度学习是机器学习的最新突破方向,它在过去的神经网络基础上找到了新的突破点,将神经网络的不足弥补上,再配合计算机硬件的进步,实现了当前最流行的机器学习算法。本系统采用的是降噪自动编码机算法,针对手写数字图像库进行训练,实现用户上传手写数字图片识别和检索功能。通过多次试验证实,本算法较传统神经网络算法在精度上有较大提高。Deep learning is one of the new breakthroughs in the direction of machine learning. Making up the deficiencies of the neural network and absorbing the advancement of computer hardware, the author achieves the most popular machine learning algorithm. This system has adopted the noise re- duction algorithm for automatic coding machine, with the help of the training for handwritten digit images input, which is extremely friendly for users to realize the recognition and retrieval function of the uploaded handwritten digital images. A number of experiments have proven that this algorithm is better than the traditional neural network algorithm in its greater improvement in accuracy.

关 键 词：深度学习 机器学习 神经网络 自动编码机 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]