摘要
利用卷积神经网络进行实验的分析设计,实现一类图像数据的分类研究。实验过程选取参考性较高的几个开源数据集,分别应用到具体的网络模型中进行识别和准确性对比,针对得到的实验结果,进行分析和改进,并给出具体的改进说明。
Focuses on image classification of certain datasets by using a convolutional neural network, analyzes the mathematical model on how to design the experiments, chooses several open source datasets, gives the experimental results and an accuracy comparison between models followed with some analysis.
基金
国家自然科学基金资助项目(61303015)
四川省科技计划项目(No.2014GZ0005-5)
关键词
深度神经网络
图像模式识别
卷积神经网络
Deep Neural Network
Image Recognition
Convolutional Neural Network
