在影像處理中，影像超分辨率在視覺上的任務一直都佔有重要地位。從傳統演算法到現在流行的深度學習方法，然而現在的深度學習方法需要大量的參數量和計算量，在嵌入式裝置上或是需要低耗電、容量不夠的裝置下不是那麼友善，為此我們提出一個更輕量且快速的網路、我們使用U-shape網路模型架構，並用了更輕量的網路模塊，將原本的3x3卷積換成1x1卷積、以及將主要模塊裡面的淺殘差塊替換成我們自己設計的模塊、大量實驗上也說明、這個模型是能夠達到比以往模型更輕而且效果更好。我們也提出了兩個新的注意力模塊、一個注意力模塊(Enhanced Spatial Channel Attention, ESCA)是用來增強每個模塊裡不同特徵點資訊混合的策略、另外一個注意力模塊Asymmetric Non-Local Mapping Residual Block(ANMRB)則是能夠讓訓練更準確更穩定。我們的方法跟原本的基準(URnet)比較、參數量從612K減少到232K少了將近400K，在準確上雖然沒有贏過基準、但是卻已經比以往那些輕量網路來說好非常多。證明事實上不需要這麼多的參數量，也能夠達到相當不錯的水準，對於嵌入式設備來說是非常適合應用上去的。並也提供一個效能跟準確度取捨的一個新想法。

In image processing, image super-resolution has always played an important role in vision tasks, from traditional algorithms to the current popular deep learning methods. However, the current deep learning method requires a large number of parameters and calculations, which is not so friendly on embedded devices or devices that require low power consumption and insufficient capacity. We propose a lighter and faster network. We use a U-shape network model architecture, and a lighter network module. We replace the original 3x3 convolution with a 1x1 convolution and replace the shallow residual block in the main module with our designed block. We also propose two new attention modules. One attention module called Enhanced Spatial Channel Attention (ESCA) is used to enhance the information mixing strategy of different feature points in each module, and another attention module called Asymmetric Non-Local Mapping Residual Block (ANMRB) can make training more accurate and stable. As a result, compared with the original benchmark (URnet), our method reduces the number of parameters from 612K to 232K by nearly 400K. It proves that in fact, there is no need for so many parameters, and it can reach a very good level, which is very suitable for embedded devices.

[1] Kim, J., Lee, J. K., & Lee, K. M. (2016). Deeply-recursive convolutional network for image super-resolution. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1637-1645).
[2] Tai, Y., Yang, J., & Liu, X. (2017). Image super-resolution via deep recursive residual network. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3147-3155).
[3] Hui, Z., Gao, X., Yang, Y., & Wang, X. (2019, October). Lightweight image super-resolution with information multi-distillation network. In Proceedings of the 27th acm international conference on multimedia (pp. 2024-2032).
[4] Hui, Z., Wang, X., & Gao, X. (2018). Fast and accurate single image super-resolution via information distillation network. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 723-731).
[5] Liu, J., Tang, J., & Wu, G. (2020, August). Residual feature distillation network for lightweight image super-resolution. In European Conference on Computer Vision (pp. 41-55). Springer, Cham.
[6] Wang, Y., Zhao, L., Liu, L., Hu, H., & Tao, W. (2021). URNet: a U-shaped residual network for lightweight image super-resolution. Remote Sensing, 13(19), 3848.
[7] Dong, C., Loy, C. C., He, K., & Tang, X. (2014, September). Learning a deep convolutional network for image super-resolution. In European conference on computer vision (pp. 184-199). Springer, Cham.
[8] Kim, J., Lee, J. K., & Lee, K. M. (2016). Accurate image super-resolution using very deep convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1646-1654).
[9] Lai, W. S., Huang, J. B., Ahuja, N., & Yang, M. H. (2017). Deep laplacian pyramid networks for fast and accurate super-resolution. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 624-632).
[10] Tai, Y., Yang, J., Liu, X., & Xu, C. (2017). Memnet: A persistent memory network for image restoration. In Proceedings of the IEEE international conference on computer vision (pp. 4539-4547).
[11] Dong, C., Loy, C. C., & Tang, X. (2016, October). Accelerating the super-resolution convolutional neural network. In European conference on computer vision (pp. 391-407). Springer, Cham.
[12] Lim, B., Son, S., Kim, H., Nah, S., & Mu Lee, K. (2017). Enhanced deep residual networks for single image super-resolution. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops (pp. 136-144).
[13] Qiu, Y., Wang, R., Tao, D., & Cheng, J. (2019). Embedded block residual network: A recursive restoration model for single-image super-resolution. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 4180-4189).
[14] Zhang, Y., Tian, Y., Kong, Y., Zhong, B., & Fu, Y. (2018). Residual dense network for image super-resolution. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2472-2481).
[15] Ding, X., Guo, Y., Ding, G., & Han, J. (2019). Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 1911-1920).
[16] Wang, Z., Liu, D., Yang, J., Han, W., & Huang, T. (2015). Deep networks for image super-resolution with sparse prior. In Proceedings of the IEEE international conference on computer vision (pp. 370-378).
[17] Liu, J., Zhang, W., Tang, Y., Tang, J., & Wu, G. (2020). Residual feature aggregation network for image super-resolution. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 2359-2368).
[18] Wang, X., Girshick, R., Gupta, A., & He, K. (2018). Non-local neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 7794-7803).
[19] Zhu, Z., Xu, M., Bai, S., Huang, T., & Bai, X. (2019). Asymmetric non-local neural networks for semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 593-602).
[20] Agustsson, E., & Timofte, R. (2017). Ntire 2017 challenge on single image super-resolution: Dataset and study. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops (pp. 126-135).
[21] Bevilacqua, M., Roumy, A., Guillemot, C., & Alberi-Morel, M. L. (2012). Low-complexity single-image super-resolution based on nonnegative neighbor embedding.
[22] Zeyde, R., Elad, M., & Protter, M. (2010, June). On single image scale-up using sparse-representations. In International conference on curves and surfaces (pp. 711-730). Springer, Berlin, Heidelberg.
[23] P. Arbelaez, M. Maire, C. Fowlkes, J. Malik, “Contour detection and hierarchical image segmentation,” in Proceedings of the IEEE Transactions on Pattern Analysis and Machine Intelligence,2011.
[24] J. B. Huang, A. Singh, N. Ahuja, “Single image super-resolution from transformed self-exemplars,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
[25] Z. Wang, A. Bovik, H. Sheikh, E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” in Proceedings of the IEEE Transactions on Image Processing, 2004.
[26] D. Kingma, J. Ba, “Adam: A method for stochastic optimization,” in Proceedings of the International Conference on Learning Representations, 2014.
[27] Zhang, K., Zuo, W., & Zhang, L. (2018). Learning a single convolutional super-resolution network for multiple degradations. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3262-3271).
[28] Zhao, H., Shi, J., Qi, X., Wang, X., & Jia, J. (2017). Pyramid scene parsing network. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2881-2890).
[29] Canny, J. (1986). A computational approach to edge detection. IEEE Transactions on pattern analysis and machine intelligence, (6), 679-698.