Sign in required
Sign in
Reveal-Layer
Reveal-Layer replaces full-image blind decomposition with controllable, on-demand extraction. Specify a target boundary using bounding-box coordinates or direct selection to isolate an object as an independent RGBA layer, giving users and developers precise, Photoshop-grade layer control.
Occlusion-Aware Layer Decomposition Framework: RevealLayer jointly models foreground separation, background restoration, and occlusion completion for real-world scenes, decomposing a single RGB image into independently editable RGBA layers.
Region-Aware and Occlusion-Guided Mechanisms: Region-Aware Attention and an Occlusion-Guided Adapter focus each layer on its own region and infer occluded content from context.
Bounding-box-guided controllable layer decomposition:Provide a target box to specify the region to decompose. The model automatically extracts the foreground, restores the background, and completes occluded content.
Service workflow:After an image loads, the web experience calls object_detect to generate initial boxes, then uses submit_task and query_task to complete decomposition and records actual usage tokens for billing.
Controllability of Layer Decomposition: With simple bounding-box guidance, RevealLayer reliably decomposes selected objects into a background layer and independent foreground RGBA layers while accurately reconstructing complex occluded regions.
Overall Generation and Restoration Performance: For background restoration and foreground matting, the model reduces object remnants and structural artifacts while preserving complex edges, transparency, and fine-grained structure.
Controllability of Layer Decomposition
Evaluates background restoration, foreground separation, and layer editability in complex multi-object layouts.
Overall Generation and Restoration Performance
Covers core layer tasks including background reconstruction, object removal, and foreground matting.
Semantic Layer Separation: AI identifies subjects, backgrounds, and design elements so that every layer has a clear semantic role.
Intelligent Content Completion: After foreground separation, the model completes occluded background regions so each independent layer remains complete and natural.
User-Defined Layer Decomposition: Add or remove boxes to control which targets are decomposed for precise editing.
Transparent RGBA Output: Each separated layer includes a complete alpha channel for natural compositing, reuse, and further editing.