A SAGE Journal


Deep learning for fast bronze inscription image retrieval

Ziyang Wang, Wei Zheng, Youguang Chen

发布时间:2021-12-03    点击次数:

【摘要】Collections of bronze inscription images are increasing rapidly. To use these images efficiently, we proposed an effective content-based image retrieval framework using deep learning. Specifically, we extract discriminative local features for image retrieval using the activations of the convolutional neural network and binarize the extracted features for improving the efficiency of image retrieval, firstly. Then, we use the cosine metric and Euclidean metric to calculate the similarity between the query image and dataset images. The result shows that the proposed framework has an impressive accuracy.

【关键词】Bronze inscription, content-based image retrieval, cosine metric, deep learning, Euclidean metric

常用连接 :
华东师大文字中心 百度

编辑部地址: 上海 中山北路3663号 华东师范大学 理科大楼 A1607室 (邮政编码:200062)

                                 版权所有  《中国文字》季刊