Controllable Layer Decomposition for Reversible Multi-Layer Image Generation

TL;DR

This paper introduces Controllable Layer Decomposition (CLD), a novel method for precise, controllable multi-layer image separation that enhances editing flexibility and outperforms existing techniques in quality and controllability.

Contribution

The paper proposes two innovative modules, LD-DiT and MLCA, enabling fine-grained layer separation and conditional generation, along with a new benchmark for evaluation.

Findings

CLD outperforms existing methods in decomposition quality.

CLD provides superior controllability in multi-layer image editing.

Separated layers are compatible with standard design tools.

Abstract

This work presents Controllable Layer Decomposition (CLD), a method for achieving fine-grained and controllable multi-layer separation of raster images. In practical workflows, designers typically generate and edit each RGBA layer independently before compositing them into a final raster image. However, this process is irreversible: once composited, layer-level editing is no longer possible. Existing methods commonly rely on image matting and inpainting, but remain limited in controllability and segmentation precision. To address these challenges, we propose two key modules: LayerDecompose-DiT (LD-DiT), which decouples image elements into distinct layers and enables fine-grained control; and Multi-Layer Conditional Adapter (MLCA), which injects target image information into multi-layer tokens to achieve precise conditional generation. To enable a comprehensive evaluation, we build a new benchmark and introduce tailored evaluation metrics. Experimental results show that CLD consistently outperforms existing methods in both decomposition quality and controllability. Furthermore, the separated layers produced by CLD can be directly manipulated in commonly used design tools such as PowerPoint, highlighting its practical value and applicability in real-world creative workflows. Our project is available at https://monkek123king.github.io/CLD_page/.