Infrared image super-resolution reconstruction by using generative adversarial network with an attention mechanism
- PDF / 3,024,201 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 70 Downloads / 186 Views
Infrared image super-resolution reconstruction by using generative adversarial network with an attention mechanism Qing-Ming Liu 1 & Rui-Sheng Jia 1,2
&
Yan-Bo Liu 1 & Hai-Bin Sun 1,2 & Jian-Zhi Yu 1,2 & Hong-Mei Sun 1,2
Accepted: 28 September 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Due to the limitations of infrared imaging principles and imaging systems, many problems are typically encountered with collected infrared images, such as low resolution, insufficient detail information, and blurred edges. In response to these problems, a method of infrared image super-resolution reconstruction that uses recursive attention and is based on a generative adversarial network is proposed. First, according to the characteristics of low-resolution infrared images such as uniform pixel distributions, low contrast, and poor perceived quality, a deep generator structure with a recursive-attention network is designed in this article. The recursiveattention module is used to extract high-frequency information from the feature maps, suppress useless information, and enhance the expressiveness of the features, which facilitates the reconstruction of texture details of infrared images. Then, to better distinguish the reconstructed images from the original high-resolution images, we designed a discriminator that was composed of a deep convolutional neural network. In addition, targeted improvements were made to the content loss function of GAN. We used the pre-trained VGG-19 network features before activation to calculate the perceptual loss, which helps recover the texture details of the infrared images. The experimental results on infrared image datasets demonstrated that the reconstruction performance of the proposed method is higher than those of several typical methods, and it realizes higher image visual quality. Keywords Infrared images . Super-resolution reconstruction . Attention mechanism . Generative adversarial network
1 Introduction With the development of infrared sensor technology, infrared images have been widely used in military, aerospace, medical, remote sensing, and other applications [1, 2]. Due to the limitations of infrared imaging principles and hardware performance, compared with natural images, infrared images have low spatial resolution and inconspicuous contrast. The industry mainly improves the hardware performances of infrared imaging systems by improving the manufacturing processes of infrared sensors to increase the spatial resolution of the obtained infrared images. Compared overcoming the
* Rui-Sheng Jia [email protected] * Hong-Mei Sun [email protected] 1
College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao 266590, China
2
Shandong Province Key Laboratory of Wisdom Mine Information Technology, Shandong University of Science and Technology, Qingdao 266590, China
limitations of hardware, it is more economical and practical to increase the resolution of infrared images by using superresolution reconstruction (
Data Loading...
