NSSR-DIL:利用深度身份学习实现空镜头图像超分辨率

Sree Rama Vamsidhar S, Rama Krishna Gorthi
{"title":"NSSR-DIL:利用深度身份学习实现空镜头图像超分辨率","authors":"Sree Rama Vamsidhar S, Rama Krishna Gorthi","doi":"arxiv-2409.12165","DOIUrl":null,"url":null,"abstract":"The present State-of-the-Art (SotA) Image Super-Resolution (ISR) methods\nemploy Deep Learning (DL) techniques using a large amount of image data. The\nprimary limitation to extending the existing SotA ISR works for real-world\ninstances is their computational and time complexities. In this paper, contrary\nto the existing methods, we present a novel and computationally efficient ISR\nalgorithm that is independent of the image dataset to learn the ISR task. The\nproposed algorithm reformulates the ISR task from generating the Super-Resolved\n(SR) images to computing the inverse of the kernels that span the degradation\nspace. We introduce Deep Identity Learning, exploiting the identity relation\nbetween the degradation and inverse degradation models. The proposed approach\nneither relies on the ISR dataset nor on a single input low-resolution (LR)\nimage (like the self-supervised method i.e. ZSSR) to model the ISR task. Hence\nwe term our model as Null-Shot Super-Resolution Using Deep Identity Learning\n(NSSR-DIL). The proposed NSSR-DIL model requires fewer computational resources,\nat least by an order of 10, and demonstrates a competitive performance on\nbenchmark ISR datasets. Another salient aspect of our proposition is that the\nNSSR-DIL framework detours retraining the model and remains the same for\nvarying scale factors like X2, X3, and X4. This makes our highly efficient ISR\nmodel more suitable for real-world applications.","PeriodicalId":501130,"journal":{"name":"arXiv - CS - Computer Vision and Pattern Recognition","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"NSSR-DIL: Null-Shot Image Super-Resolution Using Deep Identity Learning\",\"authors\":\"Sree Rama Vamsidhar S, Rama Krishna Gorthi\",\"doi\":\"arxiv-2409.12165\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The present State-of-the-Art (SotA) Image Super-Resolution (ISR) methods\\nemploy Deep Learning (DL) techniques using a large amount of image data. The\\nprimary limitation to extending the existing SotA ISR works for real-world\\ninstances is their computational and time complexities. In this paper, contrary\\nto the existing methods, we present a novel and computationally efficient ISR\\nalgorithm that is independent of the image dataset to learn the ISR task. The\\nproposed algorithm reformulates the ISR task from generating the Super-Resolved\\n(SR) images to computing the inverse of the kernels that span the degradation\\nspace. We introduce Deep Identity Learning, exploiting the identity relation\\nbetween the degradation and inverse degradation models. The proposed approach\\nneither relies on the ISR dataset nor on a single input low-resolution (LR)\\nimage (like the self-supervised method i.e. ZSSR) to model the ISR task. Hence\\nwe term our model as Null-Shot Super-Resolution Using Deep Identity Learning\\n(NSSR-DIL). The proposed NSSR-DIL model requires fewer computational resources,\\nat least by an order of 10, and demonstrates a competitive performance on\\nbenchmark ISR datasets. Another salient aspect of our proposition is that the\\nNSSR-DIL framework detours retraining the model and remains the same for\\nvarying scale factors like X2, X3, and X4. This makes our highly efficient ISR\\nmodel more suitable for real-world applications.\",\"PeriodicalId\":501130,\"journal\":{\"name\":\"arXiv - CS - Computer Vision and Pattern Recognition\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Computer Vision and Pattern Recognition\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.12165\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Computer Vision and Pattern Recognition","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.12165","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

目前的最新(SotA)图像超分辨率(ISR)方法采用深度学习(DL)技术,使用大量图像数据。将现有的 SotA ISR 作品扩展到现实世界中的主要限制在于其计算和时间复杂性。在本文中,与现有方法相反,我们提出了一种新颖且计算效率高的 ISR 算法,它独立于图像数据集来学习 ISR 任务。所提出的算法将 ISR 任务从生成超级分辨率(SR)图像重新表述为计算跨越降解空间的核的逆。我们引入了深度身份学习(Deep Identity Learning),利用降解模型和逆降解模型之间的身份关系。所提出的方法既不依赖于 ISR 数据集,也不依赖于单一输入的低分辨率(LR)图像(如自监督方法,即 ZSSR)来为 ISR 任务建模。因此,我们将我们的模型称为使用深度身份学习的空镜头超分辨率(NSSR-DIL)。所提出的 NSSR-DIL 模型所需的计算资源更少,至少减少了 10 倍,并且在基准 ISR 数据集上表现出了极具竞争力的性能。我们主张的另一个显著特点是,NSSR-DIL 框架不需要重新训练模型,并且在 X2、X3 和 X4 等规模因子变化时保持不变。这使得我们的高效 ISR 模型更适合实际应用。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
NSSR-DIL: Null-Shot Image Super-Resolution Using Deep Identity Learning
The present State-of-the-Art (SotA) Image Super-Resolution (ISR) methods employ Deep Learning (DL) techniques using a large amount of image data. The primary limitation to extending the existing SotA ISR works for real-world instances is their computational and time complexities. In this paper, contrary to the existing methods, we present a novel and computationally efficient ISR algorithm that is independent of the image dataset to learn the ISR task. The proposed algorithm reformulates the ISR task from generating the Super-Resolved (SR) images to computing the inverse of the kernels that span the degradation space. We introduce Deep Identity Learning, exploiting the identity relation between the degradation and inverse degradation models. The proposed approach neither relies on the ISR dataset nor on a single input low-resolution (LR) image (like the self-supervised method i.e. ZSSR) to model the ISR task. Hence we term our model as Null-Shot Super-Resolution Using Deep Identity Learning (NSSR-DIL). The proposed NSSR-DIL model requires fewer computational resources, at least by an order of 10, and demonstrates a competitive performance on benchmark ISR datasets. Another salient aspect of our proposition is that the NSSR-DIL framework detours retraining the model and remains the same for varying scale factors like X2, X3, and X4. This makes our highly efficient ISR model more suitable for real-world applications.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信