Meta tags:

Headings (most frequently used words):

with, applications, additional, camera, motioncanvas, cinematic, shot, design, controllable, image, to, video, generation, abstract, comparison, showcases, comparisons, baseline, methods, ablation, study, analysis, large, motion, results, legend, of, motions,

Text of the page (most frequently used words):
motion (64), camera (53), object (42), video (31), ours (27), #motioncanvas (21), mofa (18), draganything (18), local (17), global (16), and (15), dolly (15), control (12), inputs (10), input (10), right (10), image (10), design (10), left (9), the (8), results (8), shot (8), aware (7), with (7), out (7), pedestal (7), man (6), trucking (6), generation (6), motions (5), stream (5), density (5), panning (5), static (5), applications (5), orbit (5), cinematic (5), scene (5), point (4), additional (4), zoom (4), result (4), sample (4), source (3), code (3), content (3), crossing (3), roll (3), clockwise (3), comparisons (3), editing (3), method (3), that (3), users (3), both (3), website (2), them (2), creative (2), legend (2), transformation (2), transform (2), red (2), backpack (2), steps (2), valley (2), effect (2), text (2), track (2), analysis (2), coord (2), conditioning (2), traj (2), coeff (2), plucker (2), gauss (2), map (2), ablation (2), study (2), cameractrl (2), motionctrl (2), trackdiffusion (2), transfer (2), long (2), trcuking (2), joint (2), tilting (2), effectiveness (2), manner (2), showcases (2), challenges (2), user (2), into (2), i2v (2), models (2), demonstrate (2), space (2), intentions (2), for (2), diffusion (2), paper (2), presents (2), movements (2), effectively (2), university (2), controllable (2), based, project, page, you, want, reuse, their, please, credit, appropriately, tooncrafter, nerfies, licensed, under, commons, attribution, noncommercial, sharealike, international, license, frame, version, large, preview, essentiality, transformations, turning, around, raining, wearing, blue, flannel, shirt, hiking, boots, carefully, across, rocky, picturesque, surrounded, rugged, mountains, over, mountain, prompt, show, different, levels, detail, bounding, box, representation, reference, davis, baseline, methods, signal, videos, complex, trajectories, diagonal, bottom, down, tiliting, comparison, address, these, introduce, integrates, driven, controls, allowing, connecting, insights, from, classical, computer, graphics, contemporary, techniques, ability, achieve, synthesis, without, requiring, costly, related, training, data, enables, intuitively, depict, translates, spatiotemporal, signals, our, wide, range, real, world, scenarios, highlighting, its, potential, enhance, workflows, digital, creation, adapt, various, this, allows, shots, context, critical, aspect, filmmaking, involves, meticulously, planning, however, enabling, intuitive, modern, systems, two, main, first, capturing, where, must, specified, jointly, second, representing, information, can, utilized, model, synthesize, animations, abstract, full, res, arxiv, monash, adobe, research, chinese, hong, kong, feng, liu, tien, tsin, wong, chi, wing, aniruddha, mahapatra, jiahui, huang, cusuh, ham, mai, jinbo, xing, siggraph, 2025, conference, proceedings, back, top,

Text of the page (random words):
motioncanvas cinematic shot design with controllable image to video generation back to top motioncanvas cinematic shot design with controllable image to video generation siggraph 2025 conference proceedings jinbo xing 1 long mai 2 cusuh ham 2 jiahui huang 2 aniruddha mahapatra 2 chi wing fu 1 tien tsin wong 3 feng liu 2 1 the chinese university of hong kong 2 adobe research 3 monash university arxiv code paper full res comparisons showcases for cinematic shot design applications additional applications comparisons ablation study additional analysis legend inputs camera static camera dolly out camera orbit right pedestal up object motions abstract this paper presents a method that allows users to design cinematic video shots in the context of image to video generation shot design a critical aspect of filmmaking involves meticulously planning both camera movements and object motions in a scene however enabling intuitive shot design in modern image to video generation systems presents two main challenges first effectively capturing user intentions on the motion design where both camera movements and scene space object motions must be specified jointly and second representing motion information that can be effectively utilized by a video diffusion model to synthesize the image animations to address these challenges we introduce motioncanvas a method that integrates user driven controls into image to video i2v generation models allowing users to control both object and camera motions in a scene aware manner by connecting insights from classical computer graphics and contemporary video generation techniques we demonstrate the ability to achieve 3d aware motion control in i2v synthesis without requiring costly 3d related training data motioncanvas enables users to intuitively depict scene space motion intentions and translates them into spatiotemporal motion conditioning signals for video diffusion models we demonstrate the effectiveness of our method on a wide range of real world image content and shot design scenarios highlighting its potential to enhance the creative workflows in digital content creation and adapt to various image and video editing applications comparison showcases effectiveness in cinematic shot design joint camera and object motion control in a 3d scene aware manner camera motion object global motion object local motion pedestal up dolly in draganything mofa video motioncanvas ours camera motion object global motion object local motion static draganything mofa video motioncanvas ours camera motion object global motion object local motion roll clockwise draganything mofa video motioncanvas ours camera motion object global motion object local motion tilting up draganything mofa video motioncanvas ours camera motion object global motion object local motion dolly in draganything mofa video motioncanvas ours camera motion object global motion object local motion trucking right pedestal up draganything mofa video motioncanvas ours camera motion object global motion object local motion orbit right draganything mofa video motioncanvas ours camera motion object global motion object local motion panning left tilting up draganything mofa video motioncanvas ours camera motion object global motion object local motion dolly in draganything mofa video motioncanvas ours camera motion object global motion object local motion pedestal down tiliting up draganything mofa video motioncanvas ours camera motion object global motion object local motion panning left dolly in draganything mofa video motioncanvas ours camera motion object global motion object local motion panning left draganything mofa video motioncanvas ours camera motion object global motion object local motion trucking right draganything mofa video motioncanvas ours camera motion object global motion object local motion static draganything mofa video motioncanvas ours camera motion object global motion object local motion trucking left dolly out draganything mofa video motioncanvas ours camera motion object global motion object local motion panning left pedestal up orbit right draganything mofa video motioncanvas ours applications shot design with joint camera and object control inputs camera trucking right camera zoom in camera roll clockwise inputs camera static camera dolly in camera diagonal bottom right camera dolly out camera orbit left camera pedestal up camera orbit left camera trcuking left pedestal up camera dolly in camera dolly out camera dolly in camera trcuking left long videos with complex trajectories input image motion control signal result sample 1 result sample 2 input image result sample 1 result sample 2 object local motion control inputs results inputs results inputs results inputs results additional applications motion transfer input source video transfer results video editing input video editing results comparisons with baseline methods camera motion control input camera control motionctrl cameractrl ours dolly in zoom out dolly zoom trucking right object motion control inputs draganything mofa video camera trackdiffusion ours static inputs draganything mofa video camera trackdiffusion ours trucking right camera motion control on davis reference motionctrl cameractrl ours ablation study camera motion representation input gauss map plucker point traj coeff ours dolly out panning right input gauss map plucker point traj coeff ours roll clockwise zoom out bounding box conditioning input ours coord ours input ours coord ours additional analysis effect of point track density on camera motion control density 0 1 density 0 4 density 0 7 density 1 0 input point track results effect of text prompt we show camera motion control of dolly in with different levels of text detail a man a man crossing a stream a man with a red backpack steps over a stream in a mountain valley a man wearing a blue flannel shirt hiking boots and a red backpack carefully steps across a rocky stream in a picturesque valley surrounded by rugged mountains a man crossing a stream it is raining a man crossing a stream and turning around essentiality of camera aware and camera object aware transformations inputs preview w o transform w transform ours camera aware transformation camera object aware transformation large camera motion results motioncanvas 32 frame version legend of camera motions website content is licensed under a creative commons attribution noncommercial sharealike 4 0 international license website source code based on the nerfies project page and tooncrafter if you want to reuse their source code please credit them appropriately