PrismLayers: Open Data for High-Quality Multi-Layer Transparent Image Generative Models

2025-05-28

AI Summary

Multi-layer transparent image generation is hindered by the absence of high-quality, large-scale datasets. To address this, we introduce PrismLayersPro—the first open-source, high-fidelity multi-layer transparent image dataset comprising 200K samples. We propose a training-free diffusion-based synthesis pipeline and design two complementary architectures: LayerFLUX for single-layer generation and MultiLayerFLUX for joint multi-layer synthesis—enabling the first text-driven, semantic-layout-guided, and editable layered image generation. Key innovations include precise alpha matte modeling, semantic layout conditioning, and a rigorous human curation protocol. Quantitative and user studies demonstrate that our ART+ model achieves superior user preference (60% win rate) over the original ART and matches FLUX.1-[dev] in visual fidelity. All data, models, and tooling are fully open-sourced.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Security and Privacy: Data transparency and provenance
• Responsible Web: Algorithmic accountability and transparency on the web
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Generating high-quality, multi-layer transparent images from text prompts can unlock a new level of creative control, allowing users to edit each layer as effortlessly as editing text outputs from LLMs. However, the development of multi-layer generative models lags behind that of conventional text-to-image models due to the absence of a large, high-quality corpus of multi-layer transparent data. In this paper, we address this fundamental challenge by: (i) releasing the first open, ultra-high-fidelity PrismLayers (PrismLayersPro) dataset of 200K (20K) multilayer transparent images with accurate alpha mattes, (ii) introducing a trainingfree synthesis pipeline that generates such data on demand using off-the-shelf diffusion models, and (iii) delivering a strong, open-source multi-layer generation model, ART+, which matches the aesthetics of modern text-to-image generation models. The key technical contributions include: LayerFLUX, which excels at generating high-quality single transparent layers with accurate alpha mattes, and MultiLayerFLUX, which composes multiple LayerFLUX outputs into complete images, guided by human-annotated semantic layout. To ensure higher quality, we apply a rigorous filtering stage to remove artifacts and semantic mismatches, followed by human selection. Fine-tuning the state-of-the-art ART model on our synthetic PrismLayersPro yields ART+, which outperforms the original ART in 60% of head-to-head user study comparisons and even matches the visual quality of images generated by the FLUX.1-[dev] model. We anticipate that our work will establish a solid dataset foundation for the multi-layer transparent image generation task, enabling research and applications that require precise, editable, and visually compelling layered imagery.

Problem

• Lack of high-quality multi-layer transparent image datasets
• Need for training-free synthesis of multi-layer transparent images
• Developing models matching modern text-to-image generation aesthetics

Innovation

• Released PrismLayersPro dataset with 200K transparent images
• Training-free synthesis pipeline using diffusion models
• ART+ model for multi-layer generation with LayerFLUX