Journal of Computer Applications ›› 2020, Vol. 40 ›› Issue (7): 2046-2052.DOI: 10.11772/j.issn.1001-9081.2019112041

• Virtual reality and multimedia computing • Previous Articles     Next Articles

Low-resolution image recognition algorithm with edge learning

LIU Ying1,2,3,4, LIU Yuxia1, BI Ping1,2,3,4   

  1. 1. School of Communication and Information Engineering, Xi'an University of Posts and Telecommunications, Xi'an Shaanxi 710121, China;
    2. Key Laboratory of Electronic Information Application Technology for Crime Scene Investigation, Ministry of Public Security;(Xi'an University of Posts and Telecommunications), Xi'an Shaanxi 710121, China;
    3. International Joint Research Center for Wireless Communication and Information Processing Technology of Shaanxi Province;(Xi'an University of Posts and Telecommunications), Xi'an Shaanxi 710121, China;
    4. Center for Image and Information Processing, Xi'an University of Posts and Telecommunications, Xi'an Shaanxi 710121, China
  • Received:2019-12-02 Revised:2020-02-09 Online:2020-07-10 Published:2020-06-29
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61601362), the Specially Funded Project for Basic Work of Sci-tech Strengthening Police of the Ministry of Public Security (2018GABJC39, 2019GABJC41), the Graduate Innovation Fund Project of Xi an University of Posts and Telecommunications (CXJJLD2019017).

基于边缘学习的低分辨率图像识别算法

刘颖1,2,3,4, 刘玉霞1, 毕萍1,2,3,4   

  1. 1. 西安邮电大学 通信与信息工程学院, 西安 710121;
    2. 电子信息现场勘验应用技术公安部重点实验室(西安邮电大学), 西安 710121;
    3. 无线通信与信息处理技术国际联合研究中心(西安邮电大学), 西安 710121;
    4. 西安邮电大学 图像与信息处理研究所, 西安 710121
  • 通讯作者: 毕萍
  • 作者简介:刘颖(1972-),女,陕西户县人,高级工程师,博士,主要研究方向:图像检索、图像与视频的理解与识别;刘玉霞(1995-),女,陕西绥德人,硕士研究生,主要研究方向:图像识别、图像重建;毕萍(1981-),女,天津人,讲师,硕士,主要研究方向:模式识别、智能计算。
  • 基金资助:
    国家自然科学基金资助项目(61601362);公安部科技强警基础工作专项(2018GABJC39,2019GABJC41);西安邮电大学研究生创新基金资助项目(CXJJLD2019017)。

Abstract: Due to the influence of lighting conditions, shooting angles, transmission equipments and the surrounding environments, target objects in criminal investigation video images often have low-resolution, which are difficult to recognize. In order to improve the recognition rate of low-resolution images, on the basis of the classic LeNet-5 recognition network, a low-resolution image recognition algorithm based on adversarial edge learning was proposed. Firstly, the adversarial edge learning network was used to generate the fantasy edge of low-resolution image, which is similar to the edge of high-resolution image. Secondly, the edge information of this low-resolution image was fused into the recognition network as prior information for the recognition of the low-resolution image. Experiments were performed on three datasets:MNIST, EMNIST and Fashion-mnist. The results show that fusing the fantasy edge of low-resolution image into the recognition network can effectively increase the recognition rate of low-resolution images.

Key words: image recognition, low-resolution image, adversarial edge learning, generative adversarial network, LeNet-5 recognition network

摘要: 由于受光照条件、拍摄角度、传输设备以及周围环境的影响,刑侦视频图像中的目标物体往往分辨率较低,难以识别。针对低分辨率图像识别问题,在经典LeNet-5识别网络的基础上,提出了一种基于边缘学习的低分辨率图像识别算法。首先由边缘生成对抗网络生成低分辨率图像的幻想边缘,该边缘与高分辨率图像边缘相近;再将该低分辨图像的生成边缘信息作为先验信息融合到识别网络中对低分辨率图像进行识别。在MNIST、EMNIST和Fashion-mnist三个数据集上分别进行实验,结果表明,将低分辨图像的幻想边缘信息融合到识别网络中可以提高低分辨率图像的识别率。

关键词: 图像识别, 低分辨率图像, 对抗性边缘学习, 生成对抗网络, LeNet-5识别网络

CLC Number: