arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2007.04344eess.IVcs.CV
收藏

轻量级图像超分辨率,具有增强的CNN

Lightweight image super-resolution with enhanced CNN

Harbin Institute of Technology, Shenzhen(哈尔滨工业大学(深圳));Shenzhen Key Laboratory of Visual Object Detection and Recognition(深圳市视觉目标检测与识别重点实验室);Peng Cheng Laboratory(鹏城实验室);Harbin Institute of Technology(哈尔滨工业大学);University of North Carolina at Charlotte(北卡罗来纳大学夏洛特分校);National Tsing Hua University(台湾清华大学)

Chunwei Tian, Ruibin Zhuge, Zhihao Wu, Yong Xu, Wangmeng Zuo, Chen Chen, Chia-Wen Lin

摘要:

具有强大表达能力的深度卷积神经网络(CNN)在单图像超分辨率(SISR)上取得了令人印象深刻的性能。然而,它们的大量卷积和参数通常消耗高的计算成本和更多的内存存储来训练SR模型,这限制了它们在现实世界中资源受限设备的SR的应用。为了解决这些问题,我们提出了一个轻量级的增强SR CNN(LESR CNN)与三个连续的子块,信息提取和增强块(IEEB),重建块(RB)和信息细化块(IRB)。具体地说,IEEB提取分层低分辨率(LR)的功能和聚合所获得的功能逐步增加的记忆能力的浅层深层SISR。为了去除冗余信息,IEEB采用了异构体系结构。在此基础上,RB通过融合全局和局部特征,将低频特征转化为高频特征,与IEEB在解决长期依赖问题上形成互补。最后,IRB使用来自RB的粗略高频特征来学习更准确的SR特征并构建SR图像。所提出的LESRCNN可以通过不同尺度的模型获得高质量的图像。大量的实验表明,所提出的LESRCNN优于国家的最先进的SISR在定性和定量评价。LESRCNN的代码可以在https://github.com/hellloxiaotian/LESRCNN上访问。

英文摘要:

Deep convolutional neural networks (CNNs) with strong expressive ability have achieved impressive performances on single image super-resolution (SISR). However, their excessive amounts of convolutions and parameters usually consume high computational cost and more memory storage for training a SR model, which limits their applications to SR with resource-constrained devices in real world. To resolve these problems, we propose a lightweight enhanced SR CNN (LESRCNN) with three successive sub-blocks, an information extraction and enhancement block (IEEB), a reconstruction block (RB) and an information refinement block (IRB). Specifically, the IEEB extracts hierarchical low-resolution (LR) features and aggregates the obtained features step-by-step to increase the memory ability of the shallow layers on deep layers for SISR. To remove redundant information obtained, a heterogeneous architecture is adopted in the IEEB. After that, the RB converts low-frequency features into high-frequency features by fusing global and local features, which is complementary with the IEEB in tackling the long-term dependency problem. Finally, the IRB uses coarse high-frequency features from the RB to learn more accurate SR features and construct a SR image. The proposed LESRCNN can obtain a high-quality image by a model for different scales. Extensive experiments demonstrate that the proposed LESRCNN outperforms state-of-the-arts on SISR in terms of qualitative and quantitative evaluation. The code of LESRCNN is accessible on https://github.com/hellloxiaotian/LESRCNN.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑