摘要
为了实现更准确、更高效的图像信息检索,提出了基于深度神经网络的图像信息快速检索方法。建立了基于深度神经网络模型的图像信息快速检索目标函数,并设计了图像信息快速检索方法,将量化误差、平均准确率以及检索时间作为实验的测试指标,分别对所设计的检索方法以及其他3种方法进行对比分析。实验结果显示,所设计的检索方法的检索损失误差更小,检索平均准确率更高,且检索时间更短。
To achieve more accurate and efficient image information retrieval,a fast image information retrieval method based on deep neural network is proposed.The object function of fast image information retrieval based on deep neural network model is established to realize the design of fast image information retrieval method.Taking quantization error,average accuracy and retrieval time as the test indicators of the experiment,the designed retrieval methods and other three methods were compared and analyzed.The results show that the retrieval loss error of this method is smaller,the average retrieval accuracy is higher,and the retrieval time is shorter.
作者
张梦君
ZHANG Mengjun(Department of Information,Minbei Vocational and Technical College,Nanping 353000,China)
出处
《新乡学院学报》
2022年第12期19-22,共4页
Journal of Xinxiang University
关键词
深度神经网络
图像信息
快速检索
图像数据集
deep neural network
image information
quick retrieval
image data set
