Lightweight image deblurring method based on recursive gated convolution

The invention discloses a lightweight image deblurring method based on recursive gating convolution. The method comprises the following steps: acquiring a target blurred image; inputting the target blurred image into the trained lightweight image deblurring model to obtain a clear deblurred image corresponding to the target blurred image; the lightweight image deblurring model comprises a feature coding module, a feature processing module and a feature decoding module. Wherein the feature processing module is formed by connecting a plurality of self-attention modules and a recursive gating selection module in series. According to the lightweight image deblurring method based on recursive gating convolution disclosed by the invention, the advantages of Transform and the convolutional neural network can be taken into account, and a high-quality deblurred image can be obtained at relatively low time-space cost. 本发明公开了一种基于递归门控卷积的轻量级图像去模糊方法，包括：获取目标模糊图像；将目标模糊图像输入至训练好的轻量级图像去模糊模型中，得到目标模糊图像对应的清晰去模糊图像；其中，轻量级图像去模糊模型包括：特