[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83182-en":3,"doc-seo-83182-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},83182,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Making Implicit Preservation Intent Explicit in Conversational Image Editing","Conversational image editing must preserve both visible content and regions that disappear temporarily across dialogue turns. When added or modified objects occlude previously visible areas, the system should restore the original appearance if the content was never semantically changed. Existing models often fail to recover such occluded-but-unchanged content, leading to inconsistent or hallucinated results. The work introduces OCCUR-Bench, a diagnostic benchmark for temporal preservation with occlusion-and-revelation scenarios and restoration references. It also proposes ReSpec, a training-free method that turns implicit preservation into explicit restoration instructions by leveraging editing history to guide an in-context editor.","Making Implicit Preservation Intent Explicit in  \nConversational Image Editing  \nSoomin Han1,†, Jihyung Ahn2,†, Bumsoo Kim3,* , Buru Chang2,*  \n1 Sogang University, 2 Korea University, 3 Chung-Ang University  \n[soominsion@sogang.ac.kr](soominsion@sogang.ac.kr), [gina0520@korea.ac.kr](gina0520@korea.ac.kr), [bumsoo@cau.ac.kr](bumsoo@cau.ac.kr), [buru_chang@korea.ac.kr](buru_chang@korea.ac.kr)  \n†Equal contribution. * Corresponding authors.  \narXiv :2607 .0705 1v 1 [ cs .CV] 8 Jul 2026  \nAbstract  \nConversational image editing requires preserving not only visible content, but also content that temporarily disappears across turns. When newly added or modified content occludes a previously visible region, that region should reappear if it was never semantically changed. However, existing systems often fail to recover such occluded-but-unchanged content, producing inconsistent or hallucinated results. We introduce OCCUR-Bench, a diagnostic benchmark for temporal preservation in conversational image editing. OCCUR-Bench provides diverse occlusion-and-revelation scenarios with historical restoration references, enabling evaluation of faithful restoration rather than plausible regeneration. We also propose ReSpec, a training-free framework that makes implicit preservation explicit by pairing restoration-aware instructions with historical visual references. Given an editing history, ReSpec identifies what should persist, selects the historical image state that provides missing visual evidence, and conditions an in-context editor on the resulting instruction and reference image. Experiments show that ReSpec improves restoration fidelity and temporal consistency on OCCUR-Bench, highlighting the need to ground preservation in editing history rather than only the current image. The dataset and code are available at [https://](https://)[ ](https://)[github.com/anonymous745961852-cloud/](github.com/anonymous745961852-cloud/)[ ](github.com/anonymous745961852-cloud/)[implicit-preservation-editing](implicit-preservation-editing.)[.](implicit-preservation-editing.)  \n1 Introduction  \nConversational image editing enables users to refine visual content through iterative dialogue. At each turn, the user requests an edit, observes the updated image, and provides a follow-up instruction. Unlike single-turn editing, each instruction must therefore be interpreted in the context of prior edits, intermediate results, and user expectations accumulated across turns. A basic expectation is  \npreservation: content that the user has not asked to change should remain consistent.  \nExisting image editing methods are primarily designed to preserve what remains visible in the current image. However, conversational editing also requires preserving what temporarily disappears. Such cases naturally arise in iterative editing: common operations such as adding, moving, resizing, replacing, or stylizing objects can temporarily hide previously visible content and reveal it again in later turns. As illustrated in Figure 1, flowers added over an owl’s chest can hide the original feather texture. When the flowers are later removed or resized, the hidden texture should reappear because it was never semantically modified. Current conversational editing systems can fail in this setting, producing inconsistent or hallucinated content instead. Once the texture is occluded, the current image no longer provides visual evidence of its original appearance. This reveals a simple but important limitation: visual absence does not imply semantic change.  \nThis failure mode is not well captured by existing conversational image editing benchmarks. Most evaluations focus on instruction following, visible-region preservation, or final-output quality. They therefore do not directly test whether a model can recover content that was visible in an earlier turn, temporarily occluded, and later revealed. Asa result, a model may appear successful under standard evaluation while still failing to p","cbCaimk0HJOQotD9","https://ap.wps.com/l/cbCaimk0HJOQotD9","pdf",4585574,1,19,"English","en",105,"# Abstract\n# Introduction\n## Problem: Temporal preservation across dialogue turns\n## Limitation of existing benchmarks\n## OCCUR-Bench: diagnostic benchmark for occlusion-and-revelation\n## ReSpec: making implicit preservation explicit","[{\"question\":\"What is the core challenge in conversational image editing addressed by the paper?\",\"answer\":\"The paper focuses on preserving content that becomes temporarily occluded across dialogue turns and should reappear unchanged when revealed later.\"},{\"question\":\"Why do existing conversational image editing systems fail in this scenario?\",\"answer\":\"They rely on the current visible image evidence and the explicit instruction at each turn; when content is occluded, the model lacks visual proof of the original appearance and may hallucinate.\"},{\"question\":\"What are OCCUR-Bench and ReSpec, and how do they help?\",\"answer\":\"OCCUR-Bench evaluates temporal preservation using occlusion-and-revelation scenarios with historical restoration references. ReSpec is a training-free framework that uses editing history to generate restoration-aware instructions and reference images so the editor can recover occluded-but-unchanged regions faithfully.\"}]",1784185816,48,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"making-implicit-preservation-intent-explicit-in-conversational-image-editing","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/making-implicit-preservation-intent-explicit-in-conversational-image-editing/83182/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the core challenge in conversational image editing addressed by the paper?","Question",{"text":75,"@type":76},"The paper focuses on preserving content that becomes temporarily occluded across dialogue turns and should reappear unchanged when revealed later.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why do existing conversational image editing systems fail in this scenario?",{"text":80,"@type":76},"They rely on the current visible image evidence and the explicit instruction at each turn; when content is occluded, the model lacks visual proof of the original appearance and may hallucinate.",{"name":82,"@type":73,"acceptedAnswer":83},"What are OCCUR-Bench and ReSpec, and how do they help?",{"text":84,"@type":76},"OCCUR-Bench evaluates temporal preservation using occlusion-and-revelation scenarios with historical restoration references. ReSpec is a training-free framework that uses editing history to generate restoration-aware instructions and reference images so the editor can recover occluded-but-unchanged regions faithfully.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]