Context-Aware Feature Integration for Enhanced Fine-Grained Understanding in Vision-Language Models

Context-Aware Feature Integration for Enhanced Fine-Grained Understanding in Vision-Language Models | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 2 March 2026 V1 Latest version Share on Context-Aware Feature Integration for Enhanced Fine-Grained Understanding in Vision-Language Models Authors : Linyu Meng 0009-0004-5368-2620 [email protected] and Zihan Pei Authors Info & Affiliations https://doi.org/10.22541/au.177247934.47230024/v1 164 views 102 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Current Vision-Language Models often fall short in fine-grained visual understanding and complex multimodal reasoning, particularly for precise attribute recognition, relational understanding, and multi-step inference. To address this, we propose CAFI, a novel, lightweight, and plugand-play Context-Aware Feature Integration module for pre-trained VLM backbones. CAFI employs lightweight Transformer layers, sophisticated attention, and dynamic gating to adaptively fuse visual and linguistic features based on contextual relevance. Our approach utilizes a computationally efficient fine-tuning strategy, focusing training on CAFI while freezing most base VLM parameters. Experiments across diverse benchmarks demonstrate CAFI's effectiveness. It consistently boosts descriptive capabilities on COCO Dense Captioning, enhances reasoning on Visual Question Answering, and yields robust gains on general VQA. Ablation studies confirm the critical roles of dynamic gating and cross-attention. CAFI adds minimal parameters and negligible inference overhead, affirming its cost-effectiveness and practical utility. Qualitative analysis and human evaluations further corroborate that CAFI-enhanced models produce richer, more accurate, and contextually nuanced responses. Supplementary Material File (cafi.pdf) Download 1.79 MB Information & Authors Information Version history V1 Version 1 02 March 2026 Copyright This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License Keywords attention context-aware feature integration multimodal reasoning vision-language models Authors Affiliations Linyu Meng 0009-0004-5368-2620 [email protected] Guizhou Minzu University View all articles by this author Zihan Pei Guizhou Minzu University View all articles by this author Metrics & Citations Metrics Article Usage 164 views 102 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Linyu Meng, Zihan Pei. Context-Aware Feature Integration for Enhanced Fine-Grained Understanding in Vision-Language Models. Authorea . 02 March 2026. DOI: https://doi.org/10.22541/au.177247934.47230024/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177247934.47230024/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe11af0aff81640',t:'MTc3OTE3Mjc3NA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();