摘要
本文根据深度学习模型智能化的特点,提出了一种基于Caffe框架的深度学习缺陷检测模型,该模型的创新点主要表现在使用Dropout函数在图像特征抽象过程中,不断地踢除提取到的一般特征点,保留有效特征点,从而提高模型的分类识别率。实验结果表明,该方法的分类识别率高达97. 66%。与传统深度学习标签缺陷检测算法相比,该研究更加注重图像有效特征的提取,很大程度上提高了模型检测准确率,同时无需进行复杂的模板制作,实现方法简单,适应性强。
According to the characteristics of the deep learning model,this paper proposes a deep learning defect detection model based on Caffe framework.The innovation of this model is mainly in the use of Dropout function in the process of image feature abstraction,which is continuously kicked out and extracted.The general feature points retain valid feature points,thereby improving the classification recognition rate of the model.The experimental results show that the classification recognition rate of this method is as high as 97.66%.Compared with the traditional deep learning label defect detection algorithm,this research pays more attention to the extraction of effective features of the image,which greatly improves the accuracy of model detection,and does not require complicated template production.The implementation method is simple and adaptable.
作者
李培秀
李致金
韩可
朱超
LI Pei-xiu;LI Zhi-jin;HAN Ke;ZHU Chao(Department of Electronics and Communication,Nanjing University of Information Science and Technology,Nanjing 210044,China)
出处
《中国电子科学研究院学报》
北大核心
2019年第2期118-122,共5页
Journal of China Academy of Electronics and Information Technology
基金
国家自然科学基金(41575155)
关键词
Caffe框架
深度学习
标签缺陷
人工智能
卷积神经网络
图像分类
Caffe framework
Deep learning
Label defects
Artificial intelligence
Convolutional neural network
Image classification
