If you are not sure if the website you would like to visit is secure, you can verify it here. Enter the website address of the page and see parts of its content and the thumbnail images on this site. None (if any) dangerous scripts on the referenced page will be executed. Additionally, if the selected site contains subpages, you can verify it (review) in batches containing 5 pages.
favicon.ico: dreamvideo-omni.github.io - DreamVideo-Omni.

site address: dreamvideo-omni.github.io redirected to: dreamvideo-omni.github.io

site title: DreamVideo-Omni

Our opinion (on Monday 13 July 2026 17:27:43 UTC):

GREEN status (no comments) - no comments
After content analysis of this website we propose the following hashtags:



Meta tags:
description=DreamVideo-Omni;

Headings (most frequently used words):

the, and, omni, of, in, reference, trajectories, is, dreamvideo, image, subject, video, customization, wearing, on, with, motion, control, white, multi, man, little, up, to, hands, woman, background, ball, girl, front, identity, an, for, frame, generated, their, black, bounding, boxes, soccer, interacting, puppy, its, he, latent, comparison, first, they, smiling, move, camera, shirt, brown, green, grass, yellow, orange, pants, out, food, person, kitchen, his, it, flowers, learning, framework, stage, reward, single, object, scene, by, light, gently, both, view, holding, top, two, lens, shot, boy, field, park, jacket, sitting, stands, stretches, paws, waiting, ears, very, excited, dreamvideo2, her, from, surprise, vace, phantom, eyes, seemingly, then, tora, wan, had, skin, red, controlled, reinforcement, abstract, overall, overview, textbf, introduces, all, one, dit, that, incorporates, images, bboxes, further, enhances, fidelity, via, proposed, feedback, mechanism, which, utilizes, model, directly, evaluate, intermediate, latents, completely, bypassing, expensive, vae, decoder, faster, training, harmonious, joint, emergent, capabilities, unlocking, generation, conditioned, trajectory, engaging, pleasant, conversation, another, are, occasionally, gesturing, mouths, naturally, vividly, simulating, real, life, exchange, entire, illuminated, meticulously, crafted, cinematic, lighting, creating, warm, intimate, atmosphere, soft, falls, upon, them, movement, half, body, foreign, taking, selfie, outdoors, male, hat, female, carrying, bag, cityscape, including, bridges, buildings, moves, head, playing, sports, uniform, intently, chasing, after, cute, dog, happily, running, sunlit, grassy, stand, tongue, looking, lively, manipulating, transparent, crystal, seaside, turtleneck, sweater, pink, rabbit, headband, expression, changes, confusion, finally, back, throughout, angle, remains, unchanged, focusing, face, modern, focused, chopping, vegetables, while, busy, around, featuring, cabinets, decorative, tiles, walls, at, beginning, has, closed, deep, thought, or, feeling, opens, starts, play, tosses, catches, heartwarming, some, looks, close, people, exchanging, were, clothes, left, dark, right, held, exchanged, panoramic, cartoon, character, climbing, stairs, flat, was, pair, shoes, backpack, hair, wears, glasses, simple, wall, door,

Text of the page (most frequently used words):
and (53), the (51), omni (22), motion (19), #subject (16), dreamvideo (14), #control (14), multi (13), video (13), with (13), identity (13), customization (12), reference (12), trajectories (12), image (11), wearing (10), white (8), latent (7), little (7), man (7), for (6), university (6), background (5), front (5), hands (5), girl (5), ball (5), woman (5), stage (5), this (4), from (4), learning (4), liu (4), zhang (4), frame (4), first (4), their (4), interacting (4), puppy (4), black (4), its (4), soccer (4), camera (4), bounding (4), boxes (4), generated (4), reward (4), training (4), shirt (3), green (3), pants (3), brown (3), his (3), two (3), flowers (3), they (3), person (3), yellow (3), move (3), orange (3), grass (3), food (3), out (3), smiling (3), comparison (3), kitchen (3), both (3), framework (3), precise (3), code (2), webpage (2), project (2), controlled (2), reinforcement (2), author (2), wei (2), yujie (2), xinyu (2), shiwei (2), yuan (2), hangjie (2), xing (2), jinbo (2), chen (2), zhekai (2), wang (2), xiang (2), qiu (2), haonan (2), zhao (2), rui (2), feng (2), yutong (2), chu (2), ruihang (2), yingya (2), guo (2), yike (2), xihui (2), shan (2), hongming (2), arxiv (2), view (2), shot (2), red (2), had (2), skin (2), then (2), wan (2), tora (2), scene (2), park (2), jacket (2), sitting (2), holding (2), her (2), very (2), excited (2), stands (2), stretches (2), paws (2), seemingly (2), waiting (2), object (2), eyes (2), gently (2), single (2), phantom (2), vace (2), while (2), remains (2), surprise (2), ears (2), dreamvideo2 (2), top (2), field (2), joint (2), boy (2), light (2), lens (2), upon (2), harmonious (2), via (2), feedback (2), model (2), large (2), scale (2), diffusion (2), granularity (2), ambiguity (2), degradation (2), performance (2), preservation (2), paradigm (2), comprehensive (2), signals (2), global (2), controllability (2), introduce (2), aware (2), group (2), hong (2), kong (2), total, clicks, source, borrowed, nerfies, article, wei2026dreamvideo_omni, title, journal, preprint, 2603, 12257, year, 2026, panoramic, cartoon, character, climbing, stairs, flat, was, pair, shoes, backpack, hair, wears, glasses, simple, wall, door, close, people, exchanging, were, clothes, left, dark, right, held, exchanged, unlocking, generation, conditioned, trajectory, emergent, capabilities, heartwarming, some, looks, beginning, has, closed, deep, thought, feeling, opens, starts, play, tosses, catches, modern, focused, featuring, cabinets, decorative, tiles, walls, busy, around, chopping, vegetables, throughout, angle, unchanged, focusing, face, expression, changes, confusion, finally, back, pink, rabbit, headband, turtleneck, sweater, manipulating, seaside, transparent, crystal, cute, happily, stand, tongue, looking, lively, sunlit, grassy, running, dog, sports, uniform, intently, after, chasing, playing, half, body, foreign, male, hat, female, carrying, bag, cityscape, including, bridges, buildings, head, moves, taking, selfie, outdoors, movement, engaging, pleasant, another, are, occasionally, naturally, vividly, simulating, real, life, exchange, entire, illuminated, meticulously, crafted, cinematic, lighting, creating, warm, intimate, atmosphere, falls, them, soft, mouths, gesturing, conversation, overview, textbf, introduces, all, one, dit, that, incorporates, images, bboxes, further, enhances, fidelity, proposed, mechanism, which, utilizes, directly, evaluate, intermediate, latents, completely, bypassing, expensive, vae, decoder, faster, overall, models, have, revolutionized, synthesis, achieving, over, significant, challenge, recent, attempts, bridge, gap, often, suffer, limited, leading, suboptimal, work, present, unified, enabling, progressive, integrate, encompassing, appearances, local, dynamics, movements, ensure, robust, condition, rotary, positional, embedding, coordinate, heterogeneous, inputs, hierarchical, injection, strategy, enhance, guidance, furthermore, resolve, role, embeddings, explicitly, anchor, specific, identities, effectively, disentangling, complex, scenes, into, independent, controllable, instances, second, mitigate, design, pretrained, backbone, provides, rewards, space, prioritizing, aligned, human, preferences, supported, our, curated, dataset, dreamomni, bench, evaluation, dreamvideoomni, demonstrates, superior, generating, highquality, videos, abstract, paper, corresponding, leader, equal, contribution, national, singapore, nanyang, technological, mmlab, zhejiang, tongyi, lab, alibaba, science, technology, fudan,


Text of the page (random words):
dreamvideo omni dreamvideo omni omni motion controlled multi subject video customization with latent identity reinforcement learning yujie wei 1 xinyu liu 2 shiwei zhang 3 hangjie yuan 4 jinbo xing 3 zhekai chen 5 xiang wang 3 haonan qiu 6 rui zhao 7 yutong feng 3 ruihang chu 3 yingya zhang 3 yike guo 2 xihui liu 5 hongming shan 1 1 fudan university 2 the hong kong university of science and technology 3 tongyi lab alibaba group 4 zhejiang university 5 mmlab the university of hong kong 6 nanyang technological university 7 national university of singapore equal contribution project leader corresponding author paper code abstract while large scale diffusion models have revolutionized video synthesis achieving precise control over both multi subject identity and multi granularity motion remains a significant challenge recent attempts to bridge this gap often suffer from limited motion granularity control ambiguity and identity degradation leading to suboptimal performance on identity preservation and motion control in this work we present dreamvideo omni a unified framework enabling harmonious multi subject customization with omni motion control via a progressive two stage training paradigm in the first stage we integrate comprehensive control signals for joint training encompassing subject appearances global motion local dynamics and camera movements to ensure robust and precise controllability we introduce a condition aware 3d rotary positional embedding to coordinate heterogeneous inputs and a hierarchical motion injection strategy to enhance global motion guidance furthermore to resolve multi subject ambiguity we introduce group and role embeddings to explicitly anchor motion signals to specific identities effectively disentangling complex scenes into independent controllable instances in the second stage to mitigate identity degradation we design a latent identity reward feedback learning paradigm by training a latent identity reward model upon a pretrained video diffusion backbone this provides motion aware identity rewards in the latent space prioritizing identity preservation aligned with human preferences supported by our curated large scale dataset and the comprehensive dreamomni bench for multi subject and omni motion control evaluation dreamvideoomni demonstrates superior performance in generating highquality videos with precise controllability overall framework of dreamvideo omni overview of textbf dreamvideo omni in stage 1 the framework introduces an all in one video dit that incorporates reference images bboxes and trajectories for multi subject customization and omni motion control stage 2 further enhances identity fidelity via the proposed latent identity reward feedback learning mechanism which utilizes a latent identity reward model to directly evaluate intermediate latents completely bypassing the expensive vae decoder for faster training harmonious multi subject customization and omni motion control multi subject customization reference image 1 reference image 2 generated video a man engaging in a pleasant conversation with another man they are smiling and occasionally gesturing with their hands their mouths move naturally vividly simulating a real life exchange the entire scene is illuminated by meticulously crafted cinematic lighting creating a warm and intimate atmosphere the soft light gently falls upon both of them motion control camera movement trajectories generated video half body view of a foreign white man and woman taking selfie outdoors male wearing a light shirt a black hat holding a camera female wearing a brown top carrying a bag two smiling the background is a cityscape including bridges and buildings the lens moves and the video is shot in a head up lens subject customization omni motion control reference image bounding boxes trajectories generated video a little boy playing soccer on a green grass field the boy is wearing a yellow sports uniform and is intently chasing after a soccer ball multi subject customization omni motion control reference image bounding boxes trajectories generated video a little girl interacting with a puppy in the park the little girl wearing an orange jacket and black pants is sitting on the grass the puppy stands up and stretches out its front paws waiting for the food comparison of joint subject customization and motion control reference image bounding boxes trajectories a cute little dog happily running on a sunlit grassy field its ears stand up and its tongue out looking very excited and lively dreamvideo2 dreamvideo omni reference image bounding boxes trajectories a person wearing a black top manipulating a transparent crystal ball with both hands by the seaside dreamvideo2 dreamvideo omni comparison of subject customization single subject customization reference image a woman in front of a yellow background wearing a white turtleneck sweater and a pink rabbit ears headband her expression changes from surprise to confusion and finally back to surprise throughout the video the camera angle remains unchanged focusing on the woman s face vace phantom dreamvideo omni multi subject customization reference image a man and a woman interacting in a modern kitchen the woman is focused on chopping vegetables while the man is busy around the kitchen the kitchen is featuring white cabinets and decorative tiles on the walls vace phantom dreamvideo omni comparison of motion control single object trajectories trajectories a man wearing a white t shirt interacting with a soccer ball in front of an orange background at the beginning of the video the man has his eyes closed seemingly in deep thought or feeling then he opens his eyes smiling and starts to play with the soccer ball he gently tosses the ball up and catches it and tora wan move dreamvideo omni multi object trajectories trajectories a heartwarming scene of a little girl interacting with a puppy in the park the little girl wearing an orange jacket and black pants is sitting on the grass holding some food in her hands the puppy looks very excited it stands up and stretches out its front paws seemingly waiting for the food from the little girl tora wan move dreamvideo omni emergent capabilities unlocking image to video generation and first frame conditioned trajectory control reference image first frame trajectories close up of two people exchanging flowers in front of a white background they were wearing white clothes the person on the left had dark skin and the person on the right had yellow skin they held the red flowers in their hands and then exchanged the flowers in their hands dreamvideo omni reference image first frame trajectories panoramic view of a cartoon character climbing the stairs in the flat shot he was wearing a white shirt and green pants a pair of red shoes and a brown backpack his hair is brown and he wears glasses the background is a simple white wall with a green door frame on it dreamvideo omni reference article wei2026dreamvideo_omni title dreamvideo omni omni motion controlled multi subject video customization with latent identity reinforcement learning author wei yujie and liu xinyu and zhang shiwei and yuan hangjie and xing jinbo and chen zhekai and wang xiang and qiu haonan and zhao rui and feng yutong and chu ruihang and zhang yingya and guo yike and liu xihui and shan hongming journal arxiv preprint arxiv 2603 12257 year 2026 the source code of this webpage is borrowed from the nerfies project webpage total clicks
Thumbnail images (randomly selected): * Images may be subject to copyright.GREEN status (no comments)

    No Images


    Top 50 hastags from of all verified websites.

    Supplementary Information (add-on for SEO geeks)*- See more on header.verify-www.com

    Header

    HTTP/1.1 301 Moved Permanently
    Connection close
    Content-Length 162
    Server GitHub.com
    Content-Type text/html
    Location htt????/dreamvideo-omni.github.io/
    X-GitHub-Request-Id 9686:241FDF:59625B9:5A4B10A:6A55200F
    Accept-Ranges bytes
    Age 0
    Date Mon, 13 Jul 2026 17:27:43 GMT
    Via 1.1 varnish
    X-Served-By cache-rtm-ehrd2290026-RTM
    X-Cache MISS
    X-Cache-Hits 0
    X-Timer S1783963663.475764,VS0,VE103
    Vary Accept-Encoding
    X-Fastly-Request-ID e2054468452cd1a0f47f66b451b4f5ba2f7039ae
    HTTP/2 200
    server GitHub.com
    content-type text/html; charset=utf-8
    last-modified Wed, 27 May 2026 09:22:08 GMT
    access-control-allow-origin *
    strict-transport-security max-age=31556952
    etag W/ 6a16b7c0-e65a
    expires Mon, 13 Jul 2026 17:37:43 GMT
    cache-control max-age=600
    content-encoding gzip
    x-proxy-cache MISS
    x-github-request-id 3C10:2FC1D0:8F2C61:97CB75:6A55200F
    accept-ranges bytes
    age 0
    date Mon, 13 Jul 2026 17:27:43 GMT
    via 1.1 varnish
    x-served-by cache-lcy-egml8630046-LCY
    x-cache MISS
    x-cache-hits 0
    x-timer S1783963664.607050,VS0,VE106
    vary Accept-Encoding
    x-fastly-request-id 17c249201f4678764dc577fdfdaa3247b65ed216
    content-length 9727

    Meta Tags

    title="DreamVideo-Omni"
    charset="utf-8"
    name="description" content="DreamVideo-Omni"
    property="og:title" content="DreamVideo-Omni"
    property="og:description" content="DreamVideo-Omni: Omni-Motion Controlled Multi-Subject Video Customization with Latent Identity Reinforcement Learning"
    property="og:url" content="URL OF THE WEBSITE"
    property="og:image" content="static/image/video_t1.png"
    property="og:image:width" content="2412"
    property="og:image:height" content="1394"

    Load Info

    page size9727
    load time (s)0.299429
    redirect count1
    speed download32531
    server IP 185.199.108.153
    * all occurrences of the string "http://" have been changed to "htt???/"