{"title":"WaveMixSR-V2:以更高的效率增强超分辨率","authors":"Pranav Jeevan, Neeraj Nixon, Amit Sethi","doi":"arxiv-2409.10582","DOIUrl":null,"url":null,"abstract":"Recent advancements in single image super-resolution have been predominantly\ndriven by token mixers and transformer architectures. WaveMixSR utilized the\nWaveMix architecture, employing a two-dimensional discrete wavelet transform\nfor spatial token mixing, achieving superior performance in super-resolution\ntasks with remarkable resource efficiency. In this work, we present an enhanced\nversion of the WaveMixSR architecture by (1) replacing the traditional\ntranspose convolution layer with a pixel shuffle operation and (2) implementing\na multistage design for higher resolution tasks ($4\\times$). Our experiments\ndemonstrate that our enhanced model -- WaveMixSR-V2 -- outperforms other\narchitectures in multiple super-resolution tasks, achieving state-of-the-art\nfor the BSD100 dataset, while also consuming fewer resources, exhibits higher\nparameter efficiency, lower latency and higher throughput. Our code is\navailable at https://github.com/pranavphoenix/WaveMixSR.","PeriodicalId":501289,"journal":{"name":"arXiv - EE - Image and Video Processing","volume":"18 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"WaveMixSR-V2: Enhancing Super-resolution with Higher Efficiency\",\"authors\":\"Pranav Jeevan, Neeraj Nixon, Amit Sethi\",\"doi\":\"arxiv-2409.10582\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recent advancements in single image super-resolution have been predominantly\\ndriven by token mixers and transformer architectures. WaveMixSR utilized the\\nWaveMix architecture, employing a two-dimensional discrete wavelet transform\\nfor spatial token mixing, achieving superior performance in super-resolution\\ntasks with remarkable resource efficiency. In this work, we present an enhanced\\nversion of the WaveMixSR architecture by (1) replacing the traditional\\ntranspose convolution layer with a pixel shuffle operation and (2) implementing\\na multistage design for higher resolution tasks ($4\\\\times$). Our experiments\\ndemonstrate that our enhanced model -- WaveMixSR-V2 -- outperforms other\\narchitectures in multiple super-resolution tasks, achieving state-of-the-art\\nfor the BSD100 dataset, while also consuming fewer resources, exhibits higher\\nparameter efficiency, lower latency and higher throughput. Our code is\\navailable at https://github.com/pranavphoenix/WaveMixSR.\",\"PeriodicalId\":501289,\"journal\":{\"name\":\"arXiv - EE - Image and Video Processing\",\"volume\":\"18 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - EE - Image and Video Processing\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.10582\",\"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 - EE - Image and Video Processing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.10582","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
WaveMixSR-V2: Enhancing Super-resolution with Higher Efficiency
Recent advancements in single image super-resolution have been predominantly
driven by token mixers and transformer architectures. WaveMixSR utilized the
WaveMix architecture, employing a two-dimensional discrete wavelet transform
for spatial token mixing, achieving superior performance in super-resolution
tasks with remarkable resource efficiency. In this work, we present an enhanced
version of the WaveMixSR architecture by (1) replacing the traditional
transpose convolution layer with a pixel shuffle operation and (2) implementing
a multistage design for higher resolution tasks ($4\times$). Our experiments
demonstrate that our enhanced model -- WaveMixSR-V2 -- outperforms other
architectures in multiple super-resolution tasks, achieving state-of-the-art
for the BSD100 dataset, while also consuming fewer resources, exhibits higher
parameter efficiency, lower latency and higher throughput. Our code is
available at https://github.com/pranavphoenix/WaveMixSR.