Research map / Image generation / Text-to-image diffusion
AceTone: Bridging Words and Colors for Conditional Image Grading
Tianren Ma, Mingxiang Liao, Xijin Zhang, Qixiang Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add aesthetic reward to color similarity reward (ablation) | AceTone GRPO · AceTone-Bench Transfer | Aes.: 3.29 -> 3.57 | Improved |
| add aesthetic reward to color similarity reward (ablation) | AceTone GRPO · AceTone-Bench Transfer | DeltaE: 8.93 -> 8.98 | No clear effect |
| add LPIPS reward to color reward (ablation) | AceTone GRPO · AceTone-Bench Transfer | LPIPS: 0.11 -> 0.10 | No clear effect |
| add PSNR reward to color reward (ablation) | AceTone GRPO · AceTone-Bench Transfer | PSNR: 22.51 -> 22.58 | No clear effect |
| add GRPO reinforcement learning after SFT (ablation) | AceTone · AceTone-Bench Transfer | color metrics | Improved |
| use a harder RL LUT pool (8K LUTs) for data curation (ablation) | AceTone GRPO · AceTone-Bench Transfer | generalization | Improved |
| generate LUTs with AceTone instead of diffusion editing | InstructPix2Pix and Qwen-Image-Edit · AceTone-Bench Instruct | LPIPS: 0.33 -> 0.22 | Improved |
| use AceTone instead of JarvisArt | JarvisArt · AceTone-Bench Instruct | Gemini rank: 2.72 -> 1.43 | Improved |
| use AceTone instead of SA-LUT | SA-LUT · AceTone-Bench Transfer | DeltaE: 13.10 -> 8.98 | Improved |
| use AceTone instead of SA-LUT | SA-LUT · PST-50 | LPIPS: 0.16 -> 0.09 | Improved |
| use discrete VQ tokenizer instead of continuous LUT compression | LUT compression · LUT reconstruction | DeltaE: 3.11 -> 1.38 | Improved |
| use AceTone instead of Qwen-Image-Edit in user study (IGG) | Qwen-Image-Edit · user study IGG | Pref. (%): 34.8 -> 46.9 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SA-LUT — SA-LUT: Spatial Adaptive 4D Look-Up Table for Photorealistic Style Transfer (ICCV 2025) · better on 8 of 8 shared results
- InstructPix2Pix — InstructPix2Pix: Learning To Follow Image Editing Instructions (CVPR 2023) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CLIPtone: Unsupervised Learning for Text-based Image Tone Adjustment (CVPR 2024 · similarity 0.86)
- Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization (ICLR 2026 · similarity 0.86)
- ColorConceptBench: A Benchmark for Probabilistic Color-Concept Understanding in Text-to-Image Models (NeurIPS 2026 · similarity 0.86)
- Probabilistic Prompt Adaptation for Unified Image Aesthetics and Quality Assessment (CVPR 2026 · similarity 0.85)
- Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms (NeurIPS 2024 · similarity 0.85)
- UniPercept: Towards Unified Perceptual-Level Image Understanding across Aesthetics, Quality, Structure, and Texture (ICML 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
