[ALOE] Recent activities on resource "Enhancing Ultra High Resolution Remote Sensing Imagery Analysis with ImageRAG"ALOEhttp://aloe-project.de/AloeFeeds/action/atomFeedResourceActivities?resourceId=uwaswiLaloe-noreply@dfki.uni-kl.deMartinhttp://aloe-project.de/AloeView/action/userData?userId=lzwX6p3This feed contains recent activities on resource "Enhancing Ultra High Resolution Remote Sensing Imagery Analysis with ImageRAG" in [ALOE]http://aloe-project.de/AloeFeeds/action/atomFeedResourceActivities?resourceId=uwaswiLUser 'Martin' contributed resource 'Enhancing Ultra High Resolution Remote Sensing Imagery Analysis with ImageRAG'ALOEhttp://aloe-project.de/AloeFeeds/action/atomFeedResourceActivities?resourceId=uwaswiLaloe-noreply@dfki.uni-kl.deMartinhttp://aloe-project.de/AloeView/action/userData?userId=lzwX6p3http://aloe-project.de/AloeFeeds/action/atomFeedResourceActivities?resourceId=uwaswiL:6:2024-12-06 10:37:58.0December 6, 2024 10:37:00 AM CET: User <a href="http://aloe-project.de/AloeView/action/userData?userId=lzwX6p3"><b>Martin</b></a> contributed resource <a href="http://aloe-project.de/AloeView/action/resourceDetailed?resourceId=uwaswiL"><b>Enhancing Ultra High Resolution Remote Sensing Imagery Analysis with ImageRAG</b></a> with the following metadata:<br/><br/><i>associatedDate</i>: <b></b><br/><i>creator</i>: <b>Zilun Zhang, Haozhan Shen, Tiancheng Zhao, Yuhao Wang, Bin Chen, Yuxiang Cai, Yongheng Shang, Jianwei Yin</b><br/><i>description</i>: <b>Ultra High Resolution (UHR) remote sensing imagery (RSI) (e.g. 100,000 × 100,000 pixels or more) poses a significant challenge for current Remote Sensing Multimodal Large Language Models (RSMLLMs). If choose to resize the UHR image to standard input image size, the extensive spatial and contextual information that UHR images contain will be neglected. Otherwise, the original size of these images often exceeds the token limits of standard RSMLLMs, making it difficult to process the entire image and capture long-range dependencies to answer the query based on the abundant visual context.
In this paper, we introduce ImageRAG for RS, a training-free framework to address the complexities of analyzing UHR remote sensing imagery. By transforming UHR remote sensing image analysis task to image’s long context selection task, we design an innovative image contextual retrieval mechanism based on the Retrieval-Augmented Generation (RAG) technique, denoted as ImageRAG. ImageRAG’s core innovation lies in its ability to selectively retrieve and focus on the most relevant portions of the UHR image as visual contexts that pertain to a given query. Fast path and slow path are proposed in this framework to handle this task efficiently and effectively. ImageRAG allows RSMLLMs to manage extensive context and spatial information from UHR RSI, ensuring the analysis is both accurate and efficient.</b><br/><i>language</i>: <b></b><br/><i>license</i>: <b></b><br/><i>publisher</i>: <b></b><br/><i>resourceType</i>: <b>resource-bookmark</b><br/><i>rightsHolder</i>: <b></b><br/><i>tags</i>: <b>chatwithyourmap</b><br/><i>title</i>: <b>Enhancing Ultra High Resolution Remote Sensing Imagery Analysis with ImageRAG</b><br/>