[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86117-en":3,"doc-seo-86117-105":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},86117,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","FSFVE: Few Shot Compressed Face Video Enhancement","FSFVE presents a framework for enhancing highly compressed video during videocalls under limited bandwidth. Using only about 10 face frames, the system rapidly trains an instance-specific model in under 100 seconds, either before or during the call, then improves the remaining stream by intercepting and enhancing displayed frames. The approach requires no modification to standard conferencing apps, supports real-time inference on low-compute devices via CPU-friendly design, and demonstrates significant gains in compressed face quality both objectively and perceptually.","arXiv :2607 . 11040v1 [ cs .CV] 13 Jul 2026  \nFSFVE: Few Shot Compressed Face Video Enhancement  \nVarun Ramesh Jois, Antonella DiLillo, and James Storer  \nBrandeis University  \nWaltham, MA, USA  \n{vjois,dilant,[storer](storer}@brandeis.edu)[}](storer}@brandeis.edu)[@brandeis.edu](storer}@brandeis.edu)  \nAbstract  \nVideocalling has become a popular form of communication in the world today, with many companies providing free services for it. However, there are still millions of people around the world that experience poor quality videocalls due to limitations in bandwidth. This despite, most people having the required hardware.  \nIn this paper we present a novel framework for enhancing highly compressed videocalls. We show, that with as little as 10 frames of the face, we can rapidly (in under 100 seconds) train a model to enhance that instance of the videocall. The model can be trained either prior to or during the call, enhancing the rest of the call by producing better quality video. The video conferencing application need not be modified - it can be off the shelf with our system as a layer on top that trains quickly then simply lets the video conferencing application (e.g. Zoom) run as usual, where our system intercepts and improves images before they are displayed. The model is designed to run in realtime on low-compute devices such as a typical laptop CPU. Experimentally, we show that the model significantly improves quality of compressed face video both quantitatively as well as perceptually. Code can be found at [https://github.com/varun-jois/FSFVE](https://github.com/varun-jois/FSFVE).  \nIntroduction  \nVideocalling has increased dramatically in the past few years and is enabled by the near universal adoption of smartphones. However, there are large discrepancies when it comes to internet speeds. Millions of people around the world, particularly in the developing world, experience highly compressed videocalls with an associated diminished experience.  \nTo improve received highly compressed video, there have been many solutions proposed in the computer vision literature such as denoising, deblocking, super-resolution etc. Many of these models do a great job of ameliorating compression artifacts, but almost all of these models have the same problem-they either do not run in realtime, a necessity for videocalls, or they require GPUs to run which is beyond the reach of the average consumer at the moment.  \nIn this paper, we present a novel solution to the face video enhancement problem that runs in realtime at the receiving end on typical smartphones or laptops. Specifically, we investigate the question ”Is it possible to build a small and fast model for an instance of a videocall?” If such a model exists, we could construct it either prior or during the call, enhancing the rest of the call producing better quality video. Our contributions are the following:  \n1. We present our model FewShot Compressed Face Video Enhancement (FSFVE); a model for realtime, low-compute (CPU) compressed face video enhancement that can be trained on as little as 10 frames of the subject.  \nThis is a post-peer-review, pre-copyedit version of an article published in DCC 2026 . The final authenticated version is available online at [https://doi.org/10.1109/DCC66757.2026.00021](https://doi.org/10.1109/DCC66757.2026.00021) .  \n2. We show the effectiveness of our model, both quantitatively and qualitatively, at different high compression settings.  \nRelated Work  \nRecent years have seen an explosion in research for the face restoration task. VQFR [1] restores degraded facial images by replacing low quality encoded features with high quality vectors from a learned codebook. Its parallel decoder design fuses the codebook features and input information through a texture warping module employing the deformable convolution. GFP-GAN [2] uses generative facial priors (GFP) from a pretrained face GAN to restore details without relying on explicit geometric or","cbCaimzKixZrl4NY","https://ap.wps.com/l/cbCaimzKixZrl4NY","pdf",1639356,1,10,"English","en",105,"# Abstract\n# Introduction\n# Related Work\n# FSFVE Framework","[{\"question\":\"How does FSFVE improve compressed face video during a videocall?\",\"answer\":\"FSFVE trains an instance-specific enhancement model and improves the video stream by intercepting and enhancing images before they are displayed in the conferencing app.\"},{\"question\":\"What amount of training data does FSFVE need and how quickly can it train?\",\"answer\":\"The method can be trained with as little as 10 face frames and can complete rapid training in under 100 seconds.\"},{\"question\":\"Does FSFVE require changing the videocalling application or using a GPU?\",\"answer\":\"The conferencing application need not be modified because FSFVE works as a layer on top. The model is designed to run in real time on low-compute devices such as typical laptop CPU.\"}]",1784208618,25,{"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},"fsfve-few-shot-compressed-face-video-enhancement","",{"@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/fsfve-few-shot-compressed-face-video-enhancement/86117/",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},"How does FSFVE improve compressed face video during a videocall?","Question",{"text":75,"@type":76},"FSFVE trains an instance-specific enhancement model and improves the video stream by intercepting and enhancing images before they are displayed in the conferencing app.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What amount of training data does FSFVE need and how quickly can it train?",{"text":80,"@type":76},"The method can be trained with as little as 10 face frames and can complete rapid training in under 100 seconds.",{"name":82,"@type":73,"acceptedAnswer":83},"Does FSFVE require changing the videocalling application or using a GPU?",{"text":84,"@type":76},"The conferencing application need not be modified because FSFVE works as a layer on top. The model is designed to run in real time on low-compute devices such as typical laptop CPU.","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},[93,97,101,105,110,115,120,123,128,131,134],{"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":21,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]