Search papers, labs, and topics across Lattice.
VeraRetouch, a lightweight and fully differentiable framework, is introduced for multi-task photo retouching, addressing limitations of existing approaches that rely on non-differentiable external software and suffer from high parameter redundancy. The framework utilizes a 0.5B VLM to formulate retouching plans and a fully differentiable Retouch Renderer for end-to-end pixel-level training. Trained on AetherRetouch-1M+, a newly constructed million-scale dataset, and refined with a reinforcement learning post-training strategy (DAPO-AE), VeraRetouch achieves state-of-the-art performance with a significantly smaller footprint suitable for mobile deployment.
Ditch the clunky external tools: VeraRetouch slashes model size and unlocks end-to-end training for photo retouching with a fully differentiable architecture.
Reasoning photo retouching has gained significant traction, requiring models to analyze image defects, give reasoning processes, and execute precise retouching enhancements. However, existing approaches often rely on non-differentiable external software, creating optimization barriers and suffering from high parameter redundancy and limited generalization. To address these challenges, we propose VeraRetouch, a lightweight and fully differentiable framework for multi-task photo retouching. We employ a 0.5B Vision-Language Model (VLM) as the central intelligence to formulate retouching plans based on instructions and scene semantics. Furthermore, we develop a fully differentiable Retouch Renderer that replaces external tools, enabling direct end-to-end pixel-level training through decoupled control latents for lighting, global color, and specific color adjustments. To overcome data scarcity, we introduce AetherRetouch-1M+, the first million-scale dataset for professional retouching, constructed via a new inverse degradation workflow. Furthermore, we propose DAPO-AE, a reinforcement learning post-training strategy that enhances autonomous aesthetic cognition. Extensive experiments demonstrate that VeraRetouch achieves state-of-the-art performance across multiple benchmarks while maintaining a significantly smaller footprint, enabling mobile deployment. Our code and models are publicly available at https://github.com/OpenVeraTeam/VeraRetouch.
