Explicit Layer Modeling for Video Object Insertion and Layer Decomposition

Researchers introduced TriLayer, a large-scale triplet video dataset with aligned composite, background, and foreground layers.

They also built DBL-Diffusion, a dual-branch diffusion model that learns RGB composites and RGBA foregrounds.

The framework supports two tasks: DBL-Insert for object insertion and DBL-Decompose for layer separation.

By making video layers explicit, the approach can improve realistic compositing, reuse, and foreground-background editing workflows.