[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81699-en":3,"doc-seo-81699-105":30,"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},81699,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","PRISM: Latent Composition Consistency for Single-Image Reflection Removal","Single-image reflection removal (SIRR) restores the transmission layer from a reflection-corrupted mixture, a severely ill-posed decomposition. Existing approaches in pixel space suffer from nonlinear sRGB coupling that entangles transmission and reflection and restricts generalization. PRISM (Pretrained-latent Reflection Image Separation Model) reformulates SIRR as a latent linear separation problem by learning flow-matching dynamics on a pretrained FLUX backbone, recovering transmission and reflection in one forward pass. Latent Composition Consistency and Layer Contrastive Separation losses improve robust disentanglement and semantic separation, validated on six benchmarks with strong in-the-wild generalization.","arXiv :2606 .31513v2 [ cs .CV] 10 Jul 2026  \nPRISM: Latent Composition Consistency for Single-Image Reflection Removal  \nJunseong Shin 1 and Tae Hyun Kim 1 ,2⋆  \n1 Department of Artificial Intelligence, Hanyang University, Seoul, South Korea  \n2 Department of Computer Science, Hanyang University, Seoul, South Korea  \n{junsung6140, [taehyunkim}@hanyang.ac.kr](taehyunkim}@hanyang.ac.kr)  \nAbstract. Single-image reflection removal (SIRR) seeks to recover the transmission layer from a mixture corrupted by reflections—a severely ill-posed problem. Existing methods operate in pixel space, where the nonlinear sRGB formation model entangles the two layers and limits generalization. We observe that pretrained VAE latent spaces exhibit substantially lower coherence between image layers compared to pixel space, providing a more favorable working space for decomposition. Building on this finding, we propose PRISM (Pretrained-latent Reflection Image Separation Model), which reinterprets SIRR as a latent linear separation problem. Under an approximate additive formulation in latent space, PRISM learns a flow matching velocity field on a pretrained FLUX backbone that recovers both transmission and reflection in a single forward pass. To enforce robust disentanglement, we introduce a Latent Composition Consistency (LCC) strategy that constructs synthetic mixtures by swapping reflection latents across samples and enforces consistent decomposition via a cycle loss. We further propose a Layer Contrastive Separation (LCS) loss that promotes semantic separation between layers through patch-level contrastive learning, without requiring explicit reflection targets. Experiments on six benchmarks demonstrate that PRISM consistently outperforms state-of-the-art methods by significant margins, with strong generalization to in-the-wild images. Our project page is available at [https://junsung6140.github.io/prism/](https://junsung6140.github.io/prism/) .  \nKeywords: Reflection Removal · Flow Matching · Contrastive Learning  \n1 Introduction  \nPhotographs taken through glass windows or other transparent surfaces routinely capture unwanted reflections superimposed on the scene of interest. Beyond degrading visual quality, such reflections harm downstream tasks including 3D Gaussian Splatting [39], face recognition [38], and depth estimation [2] . Singleimage reflection removal (SIRR) seeks to recover the transmission layer T from an observed mixture I = T + R, where R is the reflection component [30] . This problem is severely ill-posed, as infinitely many (T, R) pairs satisfy the equation for a given I.  \n⋆ Corresponding author  \n2 J. Shin and T.H. Kim  \nI1 R1 I2 R2  \nT D(zT + zR1) D(zT + zR2)  \n(a) (b)  \nFig. 1: (a) Cosine similarity between transmission and reflection in pixel space vs. FLUX VAE [19] latent space on 454 pairs from SIR2 [37] . All points lie below the diagonal, confirming consistently lower coherence in latent space. (b) Latent swap results: composing zˆT with reflections zˆ1R , zˆ2R separated from different images and decoding via the VAE decoder D yields realistic mixtures that preserve transmission content while swapping only the reflection, demonstrating well-disentangled decomposition.  \nPrior works address this ambiguity through hand-crafted priors [17, 21, 34], learnable residue terms [13,50], or data-driven architectures [11,14,49] . However, operating in pixel space imposes two fundamental limitations: the additive model is only a coarse approximation under nonlinear camera responses, and architectures must simultaneously capture low-level texture and high-level semantics, often resulting in entangled features. While the RAW sensor domain offers amore faithful additive model [16], it requires specialized hardware inapplicable to existing sRGB imagery.  \nWhy latent space? We pursue an orthogonal direction: lifting the decomposition into the latent space of a pretrained generative model [18, 19, 33] . As shown in Fig. 1(a), we measu","cbCaie4uJUXd5aXi","https://ap.wps.com/l/cbCaie4uJUXd5aXi","pdf",24642659,4,1,33,"English","en",105,"# Introduction\n## Problem Background: Single-Image Reflection Removal (SIRR)\n## Why Pixel Space Limitations Matter\n## Latent Space Reformulation and PRISM Overview\n## Latent Composition Consistency (LCC)\n## Layer Contrastive Separation (LCS)","[{\"question\":\"What problem does PRISM address?\",\"answer\":\"PRISM addresses single-image reflection removal, which aims to separate the transmission layer from a mixture corrupted by reflections.\"},{\"question\":\"Why does PRISM move from pixel space to latent space?\",\"answer\":\"PRISM leverages the observation that pretrained VAE latent spaces show much lower coherence between transmission and reflection than pixel space, making decomposition more workable.\"},{\"question\":\"How does PRISM ensure robust separation of transmission and reflection?\",\"answer\":\"PRISM introduces Latent Composition Consistency using latent swapping plus a cycle loss, and it adds a Layer Contrastive Separation objective based on patch-level contrastive learning to promote semantic separation without explicit reflection targets.\"}]",1784175482,83,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"prism-latent-composition-consistency-for-single-image-reflection-removal","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":20},"https://docshare.wps.com/document/prism-latent-composition-consistency-for-single-image-reflection-removal/81699/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-22","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 problem does PRISM address?","Question",{"text":75,"@type":76},"PRISM addresses single-image reflection removal, which aims to separate the transmission layer from a mixture corrupted by reflections.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why does PRISM move from pixel space to latent space?",{"text":80,"@type":76},"PRISM leverages the observation that pretrained VAE latent spaces show much lower coherence between transmission and reflection than pixel space, making decomposition more workable.",{"name":82,"@type":73,"acceptedAnswer":83},"How does PRISM ensure robust separation of transmission and reflection?",{"text":84,"@type":76},"PRISM introduces Latent Composition Consistency using latent swapping plus a cycle loss, and it adds a Layer Contrastive Separation objective based on patch-level contrastive learning to promote semantic separation without explicit reflection targets.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]