LIU Q R,WANG C,GUO F,et al. Degradation prior guided quick response code super-resolution[J]. Microelectronics & Computer,2024,41(7):18-28. doi: 10.19304/J.ISSN1000-7180.2023.0509
Citation: LIU Q R,WANG C,GUO F,et al. Degradation prior guided quick response code super-resolution[J]. Microelectronics & Computer,2024,41(7):18-28. doi: 10.19304/J.ISSN1000-7180.2023.0509

Degradation prior guided quick response code super-resolution

  • QR codes, as a graphic representation capable of storing a large amount of information, play a very important role in various industrial applications, including automation control, logistics management, quality tracing, and information exchange during transportation. High-precision recognition of QR codes is fundamental for achieving fast and accurate information exchange. However, due to the constraints imposed by the capture environment and the precision of capturing devices, low resolutions often hinder correct recognition. To address this issue, a super-resolution reconstruction algorithm for the real degraded QR code is proposed. Considering the complexity of real degradation, a degraded prior-based super-resolution algorithm is proposed. Firstly, a degraded prior information encoder is designed to extract and encode relevant information related to the degradation of image quality caused by capture environment and device limitations. Then, a degraded prior guidance module is proposed, which uses the information extracted from the encoder to guide the feature reconstruction of the main structure, including degraded feature map guidance and degraded prior guidance. Due to the lack of relevant datasets, a real degraded QR code super-resolution dataset is first constructed, consisting of 4944 pairs of low-resolution and high-resolution QR code images. Considering the slight displacement between real data pairs, a displacement-insensitive loss function is introduced to optimize the network. Experimental results demonstrate that the proposed method outperforms five classical super-resolution reconstruction algorithms in terms of Peak Signal to Noise Ratio(PSNR), Structural Similarity(SSIM), and recognition rate, indicating the superiority of the proposed method.
  • loading

Catalog

    Turn off MathJax
    Article Contents

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return