图像风格迁移是一种计算机视觉技术，它以不同的风格重新渲染和呈现图像内容，在AI艺术创作、电影、动漫和游戏等图像视频处理领域具有广泛的应用前景。基于深度学习的图像风格迁移技术极大的提高了风格化图像质量，能够实现实时任意图像风格迁移。然而，目前的研究仍然存在一些亟待解决的问题，主要包括风格化图像质量不佳和迁移速度慢问题。有些算法无法较好地保留图像的内容结构信息，导致生成的图像出现严重畸变，视觉效果不佳；有些算法通过采用复杂网络结构生成高质量图像，导致网络规模过大、参数过多，降低迁移效率。因此，本文主要研究如何生成高质量风格化图像和提高迁移效率的问题，主要研究内容和创新点如下：

(1) 针对现有算法无法较好地全局和局部风格迁移，以及空间信息丢失严重问题，提出基于多级自适应的图像风格迁移算法。首先，将内容图像和风格图像输入基于卷积神经网络构建的编码器，在编码器上增加混合注意模块，提高特征提取能力。然后，将细化的风格特征和内容特征输入到风格迁移模块，通过特征融合、迭代训练生成风格化特征。其次，将风格化特征输入与编码器镜像对称的解码器中重构出风格化图像，在编解码器之间采用跳跃连接方式，将不同层的迁移特征图进行线性叠加到解码器中完成特征解码重构，从而增强效果。最后，采用了恒等损失函数以消除图像伪影并更好的保留内容结构信息。实验结果表明，该算法可以充分融合局部与全局信息，有效改善空间信息丢失严重的问题。相比其他算法，分别在感知相似性度、内容损失和风格损失评价指标上平均改进了0.03235、0.2425和0.52。

(2) 针对卷积神经网络感受野有限和Transformer计算成本较高的问题，提出了基于CNN-Transformer风格融合模型。首先，利用卷积层（Convolutional Neural Network，CNN）提取图像的高级特征并减小尺寸，以降低处理复杂度。其次，特征随后被送入Transformer以深入分析和融合内容与风格特征，通过双分支位置风格注意模块，该模块增加了位置编码，自适应捕捉全局信息，并匹配相近风格和内容特征。此外，多级特征嵌入帮助风格迁移模块和解码器更好地保留原图结构信息和丰富风格表达。最后，使用解码器将处理好的特征映射回图像域，生成艺术化图像。实验结果表明，该算法在内容损失、风格损失和感知相似度评价指标上均有显著提升，平均改进了0.201、1.135和0.47。

本研究提出的风格迁移技术显著提升了图像迁移的质量，展现出在影视制作和艺术创作等领域的应用潜力。未来，研究可朝向模型轻量化和探索多模态风格迁移技术方面进行进一步的优化。

Image style transfer is a computer vision technique that re-renders and presents image content in different styles, with broad applications in AI art creation, film, animation, gaming, and other image and video processing domains. Deep learning-based image style transfer techniques have greatly improved the quality of stylized images and can achieve real-time arbitrary image style transfer. However, current research still faces several pressing issues, mainly including poor quality of stylized images and slow transfer speeds. Some algorithms fail to fully preserve the structural information of the image content, resulting in significant distortion and unsatisfactory visual effects in the generated images; while others, by employing complex network structures to generate high-quality images, result in excessively large network sizes and too many parameters, reducing transfer efficiency. Therefore, this paper primarily investigates how to generate high-quality stylized images and improve transfer efficiency. The main research contents and innovations are as follows:

(1) A multi-level adaptive style transfer network is proposed to address the challenges of inadequate global and local style transfer, as well as severe spatial information loss in existing algorithms. Firstly, the content and style images are fed into an encoder built on a convolutional neural network, augmented with a mixed attention module to enhance feature extraction capability. Then, refined style and content features are inputted into the style transfer module, where iterative training and feature fusion generate stylized features. Next, the stylized features are fed into a decoder symmetrically mirrored to the encoder to reconstruct the stylized image. Skip connections between the encoder and decoder linearly combine transfer feature maps from different layers to enhance feature decoding and reconstruction. Lastly, an identity loss function is employed to eliminate image artifacts and better preserve content structure information. Experimental results demonstrate that the proposed algorithm effectively integrates local and global information, mitigating severe spatial information loss. Compared to other algorithms, it achieves average improvements of 0.03235, 0.2425, and 0.52 in perceptual similarity, content loss, and style loss evaluation metrics, respectively.

(2) The proposed model addresses the issues of limited receptive fields in Convolutional Neural Networks (CNNs) and high computational costs in Transformers by introducing a CNN-Transformer style fusion model. Initially, CNNs are employed to extract high-level features from images and reduce dimensions to lower processing complexity. Subsequently, features are inputted into Transformers to deeply analyze and fuse content and style features, facilitated by a dual-branch positional style attention module that integrates positional encoding to adaptively capture global information and match similar style and content features. Additionally, multi-level feature embedding aids in preserving original image structure and enriching style expression in style transfer modules and decoders. Finally, a decoder is used to map processed features back to the image domain to generate artistic images. Experimental results demonstrate significant improvements in content loss, style loss, and perceptual similarity evaluation metrics, with average improvements of 0.201, 1.135, and 0.47, respectively.

The style transfer technology proposed in this study significantly enhances the quality of image transfer, demonstrating its potential applications in fields such as film production and artistic creation. In the future, research could focus on further optimizations towards model lightweighting and exploring multimodal style transfer techniques.

[1] 董虎胜. 卡通风格人脸图像生成研究[J]. 现代计算机, 2021, 27(27): 94-98.

[2] 纪宗杏, 贝佳, 刘润泽, 任桐炜. 基于双路视觉Transformer的图像风格迁移[J]. 北京航空航天大学学报, 2024: 1-12.

[3] 庾晨龙, 杨天, 唐贝贝. 基于双调制策略的真实感风格迁移算法[J]. 重庆工商大学学报(自然科学版), 2024: 1-10.

[4] 张彩灯, 徐杨, 莫寒, 冯明文. 融合多注意力机制的语义调整风格迁移网络[J]. 计算机工程与应用, 2024: 1-12.

[5] 时新月, 胡文瑾, 乔浪, 康文东. 融合特征位置编码和误差修正的唐卡图像风格迁移模型[J]. 计算机辅助设计与图形学学报, 2024: 1-10.

[6] 蒋亨畅, 张笃振. 保留细节特征的图像任意风格迁移[J]. 计算机系统应用, 2024, 33 (03): 118-125.

[7] 夏明桂, 田入君, 姜会钰, 董敏. 基于改进ResNet网络和迁移学习的服装图像风格识别研究[J]. 纺织工程学报, 2024, 2 (01): 12-20.

[8] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.

[9] Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, Illia Polosukhin. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017: 30-32.

[10] Karras T, Laine S, Aila T. A style-based generator architecture for generative adversarial networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 4401-4410.

[11] Karras T, Laine S, Aittala M, et al. Analyzing and improving the image quality of stylegan[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 8110-8119.

[12] Gatys L A , Ecker A S , Bethge M ,et al.Controlling Perceptual Factors in Neural Style Transfer[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2017: 1-10.

[13] Gatys L A, Ecker A S, Bethge M. Image style transfer using convolutional neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 2414-2423.

[14] 王琪, 魏纵横, 崔曼曼. 基于深度学习的传统剪纸图像风格迁移算法研究[J]. 电脑知识与技术, 2023, 19(5): 1-5．

[15] 贺希, 侯植. 基于人工神经网络风格迁移的水墨动画应用研究[J]．西北美术, 2022( 3): 90-95．

[16] 徐鹏飞, 周腾骅, 武仲科, 等. 低算力深度学习下的图像卡通风格化研究[J]. 北京师范大学学报(自然科学版), 2021, 57(6): 888-895.

[17] 艾立超. 基于风格化绘制的非真实感渲染研究[D]. 广州市: 华南理工大学, 2010: 3-5.

[18] 钱小燕. 基于图像的非真实感艺术绘制技术综述[J]. 工程图学学报, 2010, 31(01): 6-12.

[19] 张东, 唐向宏, 张少鹏, 等. 小波变换与纹理合成相结合的图像修复[J]. 中国图象图形学报, 2015(07): 882-894.

[20] Drori I，Cohen-or D，YESHURUN H. Example-based style synthesis[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2003: 1-7.

[21] Meier B J. Painterly rendering for animation[C]//Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques, 1996: 477-484.

[22] Lu J, Barnes C, Diverdi S, et.al. RealBrush: painting with examples of physical media[J]. ACM Transactions on Graphics, 2013, 32(4): 1-12.

[23] Frigo, O., Sabater, N., Delon, J., & Hellier, P. Split and match: Example-based adaptive patch sampling for unsupervised style transfer[J]. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 553-561.

[24] Wang M, Wang B, Fei Y, et al. Towards Photo Water colorization with Artistic Verisimilitude[J]. IEEE Transactions on Visualization & Computer Graphics, 2014, 20(10): 1451-1460.

[25] 钱小燕, 肖亮, 吴慧中. 模糊颜色聚类在颜色传输中的应用[J]. 计算机辅助设计与图形学学报, 2006, 18(9): 1332-1336.

[26] Hertzmann A. Painterly rendering with curved brush stokes of multiple size[C]//Proceedings of the 25th Annual Conference on Computer Graphics and Interactive Techniques, 1998: 453-460.

[27] 谢斌, 汪宁, 范有伟. 相关对齐的总变分风格迁移新模型[J]. 中国图象图形学报, 2022(2): 14.

[28] Winnemoller H, Olsen S C, Gooch B. Real-time video abstraction[J]. ACM Transactions on Graphics (TOG), 2006, 25(3): 1221-1226.

[29] Tomasi C, Manduchi R. Bilateral filtering for gray and color images[C]//Proceeding of the IEEE International Conference on Computer Vision, 1998, 98(1): 839-846.

[30] Gooch B, Reinhard E, Gooch A. Human facial illustrations Creation and psychophysical evalution[J]. ACM Transactions on Graphics, 2004, 23(1): 27-44.

[31] J. Portilla, E.P. Simoncelli, A parametric texture model based on joint statistics of complex wavelet coefficients[J]. International Journal of Computer Vision, 2000, 40(1): 49-70.

[32] Ashikhmin M. Synthesizing natural textures[C]//Proceedings of the 2001 Symposium on Interactive 3D Graphics, 2001: 217-226.

[33] 徐晓刚, 鲍虎军, 马利庄. 基于相关性原理的多样图纹理合成方法[J]. 自然科学进展, 2002, 12(6): 665-668.

[34] Han C, Risser E, Ramamoorthi R, et al. Multiscale texture synthesis[C]//Proceedings of ACM Siggraph, 2008: 1-8.

[35] H. Lee. S. Seo, S. Ryoo. And K Yoon. Lee H, Seo S, Ryoo S, et al. Directional texture transfer[C]//International Symposium on Non-photorealistic Animation & Rendering. ACM, 2010: 43-48.

[36] Semmo A, Limberger D, Kyprianidis J E, et al. Image stylization by oil paint filtering using color palettes[C]//Proceedings International Symposium on Computational Aesthetics in Graphics, Visualization, and Imaging, 2015: 149-158.

[37] Shih Y C, Paris S, Barnes C, et al. Style transfer for headshot portraits[J]. ACM Transactions on Graphics (TOG), 2014, 33(4): 1-14.

[38] Desbenoit B, Galin E, Akkouches S. Simulating and modeling lichen growth[J]. Computer Graphics Forum, 2004, 23(3): 341-350.

[39] Julesz B. Visual Pattern Discrimination[J]. Information Theory, IRE Transactions on, 1962, 8(2): 84-92.

[40] Bergen D J H J R. Pyramid-Based Texture Analysis/Synthesis[J]. Computer Graphics ACM Siggraph, 2008, 30(6): 229-238.

[41] Portilla J, Simoncelli E P. A Parametric Texture Model Based on Joint Statistics of Complex Wavelet Coefficients[J]. International Journal of Computer Vision, 2000, 40(1): 49-70.

[42] Efros A A, Leung T K. Texture synthesis by non-parametric sampling[C]//Proceedings of the Seventh IEEE International Conference on Computer Vision, 1999(2): 1033-1038.

[43] Wei L Y, Levoy M. Fast texure synthesis using tree-structured vector quantization[C]//Proceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques, 2000: 479-488.

[44] Gatys, Leon, Alexander S. Ecker, Matthias Bethge. Texture synthesis using convolutional neural networks[J]. MIT Press, 2015(28): 1-6.

[45] Gatys L A, Ecker A S, Bethge M, et al. Controlling perceptual factors in neural style transfer[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 3985-3993.

[46] Li C, Wand M. Combining markov random fields and convolutional neural networks for image synthesis[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 2479-2486.

[47] Johson J, Alahi A, FEI-FEI L. Perceptual losses for real-time style transfer and super-resolution[C]//European Conference on Computer Vision, 2016: 694-711.

[48] Ulyanov D, Lebedev V, Vedaldi A, et al. Texture networks: Feed-forward synthesis of textures and stylized images[J]. Computer Vision and Pattern Recognition, 2016: 1-5.

[49] Mirza M, Osindero S. Conditional generative adversarial nets[J]. Computer Science, 2014: 2-8.

[50] Choi Y, CHOI M, Kim M, et al. Stargan: Unified generative adversarial networks for multi-domain image-to image translation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 8789-8797.

[51] Elad, Michael, Dror Simon, and Aviad Aberdam. Another step toward demystifying deep neural networks[J]. Proceedings of the National Academy of Sciences, 2020, 117(44): 27070-27072.

[52] Nikulin, Yaroslav, and Roman Novak. Exploring the neural algorithm of artistic style[J]. Computer Science, 2016: 1-6.

[53] Novak, R., & Nikulin, Y. Improving the neural algorithm of artistic style[J]. Computer Vision and Pattern Recognition, 2016: 2-9.

[54] Risser, E., Wilmot, P., & Barnes, C. Stable and controllable neural texture synthesis and style transfer using histogram losses[J]. Computer Science, 2017: 5-10.

[55] Yin, Rujie. Content aware neural style transfer[J]. Computer Vision and Pattern Recognition, 2016: 7-10.

[56] Ho, Minh Man, J. Zhou, and Y. Fan. Respecting low-level components of content with skip connections and semantic information in image style transfer[J]. European Conference on Visual Media Production, 2019: 4-6.

[57] Liao, Yi-Sheng, and Chun-Rong Huang. Semantic context-aware image style transfer[J]. IEEE Transactions on Image Processing, 2022(31): 1911-1923.

[58] Liao, J., Yao, Y., Yuan, L., Hua, G., & Kang, S. B. Visual attribute transfer through deep image analogy[J]. ACM Transactions on Graphics, 2017: 7-11.

[59] Simonyan, K., Vedaldi, A., & Zisserman, A. Deep inside convolutional networks: Visualising image classification models and saliency maps[J]. Computer Science, 2013: 10-16.

[60] Sun, J., Sun, J., Xu, Z., Shum, H.Y.: Image super-resolution using gradient profile prior[C]//In Computer Vision and Pattern Recognition. CVPR 2008. IEEE Conference on, IEEE, 2008: 1-8.

[61] Wang X, Oxholm G, Zhang D, et al. Multimodal transfer: a hierarchical deep convolutional neural network for fast artistic style transfer[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC: IEEE Computer Society, 2017: 7178-7186.

[62] Dumoulin V, Shlens J, Kudlur M. A learned representation for artistic style[J]. Computer Science, 2016: 2-10.

[63] Li Y, Fang C, Yang J, et al. Diversified texture synthesis with feed-forward networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 3920-3928.

[64] Zhang H, Dana K. Multi-style generative network for real-time transfer[C]//Proceedings of the European Conference on Computer Vision (ECCV) Workshops, 2018: 1-5.

[65] Huang X, Belongie S. Arbitrary style transfer in real-time with adaptive instance normalization[C]//Proceedings of the IEEE International Conference on Computer Vision, 2017: 1501-1510.

[66] Li, Y., Fang, C., Yang, J., Wang, Z., Lu, X. and Yang, M.H. Universal style transfer via feature transforms[J]. Advances in Neural Information Processing Systems, 2017(30): 1-7.

[67] Li, X., Liu, S., Kautz, J., & Yang, M. H. Learning linear transformations for fast arbitrary style transfer[J]. Computer Science, 2018: 2-10.

[68] Park D Y, Lee K H. Arbitrary Style Transfer with Style-Attentional Networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019: 5880-5888.

[69] Deng Y, Tang F, Dong W, et al. Arbitrary Style Transfer via Multi-Adaptation Network[C]//2020 In Proceedings of the 28th ACM International Conference on Multimedia, 2020: 2719-2727.

[70] Yao Y, Ren J, Xie X, et al. Attention-aware Multi-stroke Style Transfer[J]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019: 1467-1475.

[71] Liu, S., Lin, T., He, D., Li, F., Wang, M., Li, X. Adaattn: Revisit attention mechanism in arbitrary neural style transfer[J]. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2020: 6649-6658.

[72] Ulyanov D, Lebedev V, Vedaldi A, et al. Improved texture networks: maximizing quality and diversity in feedforward stylization and texture synthesis[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 4105-4113.

[73] Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial networks[J]. Communications of the ACM, 2020, 63(11): 139-144.

[74] Isola P, Zhu J Y, Zhou T, et al. Image-to-image translation with conditional adversarial networks[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2017: 5967-5976.

[75] Karras T, Aittala M, Laine S, et al. Alias-free generative adversarial networks[J]. Advances in Neural Information Processing Systems, 2021, 34: 852-863.

[76] Samuth, B, Tschumperle, D, & Rabin, J. A Patch-Based Approach for Artistic Style Transfer Via Constrained Multi-Scale Image Matching[C]//IEEE International Conference on Image Processing (ICIP), IEEE, 2022: 1-5.

[77] 毛文涛, 吴桂芳, 吴超等. 基于中国写意风格迁移的动漫视频生成模型[J]. 计算机应用, 2022, 42(07): 2162-2169.

[78] 孙天鹏, 周宁宁, 黄国方. 新的基于GAN的局部写实感漫画图像风格迁移[J]. 计算机工程与应用, 2022, 58(14): 167-176.

[79] 朱仲贤, 毛语实, 蔡科伟等. 面向工业巡检的图像风格迁移方法[J]. 计算机工程与应用, 2023, 59(18): 234-241.

[80] 刘哲良, 朱玮, 袁梓洋. 结合全卷积网络与CycleGAN的图像实例风格迁移[J]. 中国图象图形学报. 2019, 24(08): 1283-1291.

[81] Yu, Yue, Ding Li, Benyuan Li, and Nengli Li. Multi-style image generation based on semantic image[J]. The Visual Computer, 2023: 1-16.

[82] Haibo Chen, Zhizhong Wang, Huiming Zhang, Zhiwen Zuo, Ailin Li, Wei Xing, Dongming Lu, et al. Artistic style transfer with internal-external learning and contrastive learning[J]. Advances in Neural Information Processing Systems, 2021: 2-6.

[83] Lin T Y, Maire M, BELONGIE S, et al. Microsoft COCO: common objects in context[C]//Computer Vision-ECCV 2014, Lecture Notes in Computer Science, 2014: 740-755.

[84] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, Attention is all you need[J]. In Advances in Neural Information Processing Systems, 2017: 3-7.

[85] Paul, S., & Chen, P. Y. Vision transformers are robust learners[J]. In Proceedings of the AAAI Conference on Artificial Intelligence. 2022, 36(2): 2071-2081.

[86] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30-35.

[87] Lin X, Xiong G, Gou G, et al. Et-bert: A contextualized datagram representation with pre-training transformers for encrypted traffic classification[J]. In Proceedings of the ACM Web Conference, 2022: 633-642.

[88] He L, Zhou Q, Li X, et al. End-to-end object detection with transformers[C]//Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020: 213-229.

[89] Deng, Y., Tang, F., Dong, W., Ma, C., Pan, X., Wang, L., Xu, C. Stytr2: Image style transfer with transformers[J]. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 11326-11336.

[90] Jaderberg M, Simonyan K, Zisserman A. Spatial transformer networks[J]. Advances in Neural Information Processing Systems, 2015, 28-29.

[91] Almahairi A, Ballas N, Cooijmans T, et al. Dynamic capacity networks[C]//International Conference on Machine Learning. PMLR, 2016: 2549-2558.

[92] Wang X, Girshick R, Gupta A, et al. Non-local neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7794-7803.

[93] Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.

[94] Li X, Wang W, Hu X, et al. Selective kernel networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 510-519.

[95] Tian Qi Chen and Mark Schmidt. Fast patch-based style transfer of arbitrary style[J]. arXiv preprint arXiv:1612.04337, 2016: 1-2.

[96] Lu Sheng, Ziyi Lin, Jing Shao, and Xiaogang Wang. Avatar-net: Multi-scale zero-shot style transfer by feature decoration[C]//In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018. 8242–8250.

[97] Woo S, Park J, Lee J Y, et al. CBAM: Convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision (ECCV), 2018: 3-19.

[98] F. Phillips and B. Mackintosh, Wikiart gallery, inc: A case for critical thinking[J]. Issues in Accounting Education, 2011, 26(3): 593-608.

[99] Deng Y, Tang F, Dong W, et al. Arbitrary Style Transfer via Multi-Adaptation Network[J]. In Proceedings of the 28th ACM International Conference on Multimedia, 2020: 2719-2727.

[100] J. An, S. Huang, Y. Song, D. Dou, W. Liu, and J. Luo, ArtFlow: Unbiased image style transfer via reversible neural flows[J]. In IEEE/CVF Conferences on Computer Vision and Pattern Recognition (CVPR), 2021, 862-871.

[101] H. Wang, Y. Zhu, B. Green, H. Adam, A. Yuille, and L.-C. Chen, Axial-deeplab: Stand-alone axial-attention for panoptic segmentation[J]. In European Conference on Computer Vision. Springer, 2020: 108-126.