Keep the person intact while motion lives inside the garment.
Garment Fill isolates upper-clothing pixels and composites a second video into that mask. The result is a hard, graphic treatment that keeps the wearer and the original frame readable.
The treatment stays recognizable while the source footage and movement change.
Garment fillPerson input example
How the effect reaches the frame.
Three stages keep the creative rule understandable and repeatable.
01
Find the garment
Human parsing identifies the upper-clothing region in the source image or video.
02
Build a firm mask
Expansion, contraction, and edge controls define exactly where the fill can appear.
03
Composite the texture
A texture video is resized and rendered only inside the clothing mask.
Implementation
Garment Video Fill
The current endpoint runs server-side on CPU and accepts hosted media URLs. Image inputs are faster because one clothing mask can be reused across the render.
Where it earns a place in the edit.
Use the effect where its visual rule clarifies a moment, not as decoration.
Fashion tests
Try moving material, graphics, and campaign footage on a look before a full post-production pass.
Music visuals
Give a performer a graphic treatment without replacing the full scene or silhouette.
Creative tooling
Call the same repeatable treatment from an internal tool or automated rendering flow.