Super-Resolution Infrared Imaging via Degraded Information Distillation Network
Yulong Zhuang, Quan Chen · 2023
Infrared imaging technology is widely used in civil and military applications. However, the cost of making high-resolution infrared detectors is expensive. For this reason, we propose an unsupervised super-resolution infrared imaging method for degraded information distillation network. We design a network model for progressive extraction of degradation information to learn more degradation information with discriminative features. We use dual-attention convolution to achieve feature adaption in both channel and space. We use sub-pixel convolution to implement the reconstruction of infrared images. We train our model using infrared images and evaluate the proposed method systematically. The experimental results show that our proposed method has good performance compared to other state-of-the-art methods.