中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
A painting authentication method based on multi-scale spatial-spectral feature fusion and convolutional neural network

文献类型:会议论文

作者Ren, Zeyue1,2,3; Wang, Huawei1,3; Kong, Fanzi1,2,3; Liu, Qing1,3; Yang, Yongqing3
出版日期2023
会议日期2023-12-15
会议地点Hybrid, Wuhan, China
关键词Painting authentication Hyperspectral imaging Multi-scale fusion Efficientnet v2
DOI10.1109/CEI60616.2023.10528061
页码395-400
英文摘要

To address the focusing challenges of varying scales at different time points of moving target images in space, we propose an improved gradient-based image sharpness evaluation method for variable focusing window. The method obtains the target focusing window adaptively and computes the sharpness evaluation value by convolving the image within the focusing window with a multi-directional gradient operator weighted factor. Experiments show that the method exhibits strong sensitivity and anti-interference capabilities, contributing to the improved performance of the space imaging autofocus system. © 2023 IEEE.

产权排序1
会议录2023 3rd International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology, CEI 2023
会议录出版者Institute of Electrical and Electronics Engineers Inc.
语种英语
ISBN号9798350372052
源URL[http://ir.opt.ac.cn/handle/181661/97511]  
专题西安光学精密机械研究所_动态光学成像研究室
通讯作者Wang, Huawei
作者单位1.Xi'an Key Laboratory of Spacecraft Optical Imaging and Measurement Technology, Xi'an, China
2.University of Chinese Academy of Sciences, Beijing, China;
3.Xi'an Institute of Optics and Precision Mechanics of Cas, Xi'an, China;
推荐引用方式
GB/T 7714
Ren, Zeyue,Wang, Huawei,Kong, Fanzi,et al. A painting authentication method based on multi-scale spatial-spectral feature fusion and convolutional neural network[C]. 见:. Hybrid, Wuhan, China. 2023-12-15.

入库方式: OAI收割

来源:西安光学精密机械研究所

浏览0
下载0
收藏0
其他版本

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。