Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 9 May 2023 (v1), revised 20 Aug 2023 (this version, v2), latest version 1 Mar 2024 (v3)]
Title:Trans-Inpainter: Wireless Channel Information- Guided Image Restoration via Multimodal Transformer
View PDFAbstract:Image inpainting is a critical computer vision task to restore missing or damaged image regions. In this paper, we propose Trans-Inpainter, a novel multimodal image inpainting method guided by Channel State Information (CSI) data. Leveraging the power of transformer architectures, Trans-Inpainter effectively extracts visual information from CSI time sequences, enabling high-quality and realistic image inpainting. To evaluate its performance, we compare Trans-Inpainter with RF-Inpainter, the state-of-the-art radio frequency (RF) signal-based image inpainting technique. Through comprehensive experiments, Trans-Inpainter consistently demonstrates superior performance in various scenarios. Additionally, we investigate the impact of CSI data variations on Trans-Inpainter's imaging ability, analyzing individual sensor data, fused data from multiple sensors, and altered CSI matrix dimensions. These insights provide valuable references for future wireless sensing and computer vision studies.
Submission history
From: Cheng Chen [view email][v1] Tue, 9 May 2023 12:26:12 UTC (10,619 KB)
[v2] Sun, 20 Aug 2023 04:02:54 UTC (8,944 KB)
[v3] Fri, 1 Mar 2024 07:42:36 UTC (17,675 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.