Research map / Image generation / Text-to-image diffusion
UniProcessor: A Text-induced Unified Low-level Image Processor
Huiyu Duan, Xiongkuo Min, Sijing Wu, Wei Shen, Guangtao Zhai
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 |
|---|---|---|---|
| text-induced context control with subject and manipulation prompts vs PromptIR | UniProcessor · CBSD68 30 degradations (average) | PSNR: 28.35 -> 30.35 | Improved |
| add Q-Former subject-aligned image representation to the context control module (ablation) | UniProcessor · Rain streak removal | PSNR: 28.23 -> 28.67 | Improved |
| add image-aware Q-Former on top of text encoder only (ablation) | UniProcessor · Rain streak removal | PSNR: 27.76 -> 28.67 | Improved |
| add context interaction module (CIM) (ablation) | UniProcessor · Low-light enhancement | PSNR: 23.03 -> 24.23 | Improved |
| inject CIM at levels 5+4+3 instead of only level 5 (ablation) | UniProcessor · Rain streak removal | PSNR: 28.30 -> 28.67 | Improved |
| add CIM at shallow levels (ablation) | UniProcessor · Rain streak and low-light | performance and compute | Mixed |
| sequential removal of both degradations with text prompts vs PromptIR twice | UniProcessor · Rain + low-light | PSNR: 23.3 -> 28.4 | Improved |
| order of sequential degradation removal | UniProcessor · Color saturate + noise | PSNR: 24.4 -> 33.9 | Mixed |
| replace swin transformer block with large-kernel conv ConvFormer block (ablation) | UniProcessor backbone · SIDD | PSNR: 39.77 -> 39.81 | Improved |
| add VQA module (LLM) to the pipeline (ablation) | UniProcessor · Compute | GMACs: 239.83 -> 678.52 | Hurt |
| UniProcessor vs PromptIR on gain size by degradation type | UniProcessor · CBSD68 JPEG vs snow | PSNR: 0.2dB for JPEG but 5.5dB for snow | Mixed |
Later papers that compared with this method
2 later accepted papers report UniProcessor as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Expert Weaving: Marrying Masked AutoRegressive and Diffusion Models for Unified Image Restoration (ECCV 2026 · better on 2 of 2 shared results)
- Learning Where and What to Restore for Composite Image Restoration (NeurIPS 2026 · better on 1 of 1 shared result)
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.
- PromptIR — PromptIR: Prompting for All-in-One Image Restoration (NeurIPS 2023) · better on 5 of 6 shared results
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 1 of 1 shared result
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
- Controlling Vision-Language Models for Multi-Task Image Restoration (ICLR 2024 · similarity 0.87)
- UNIC-Adapter: Unified Image-instruction Adapter with Multi-modal Transformer for Image Generation (CVPR 2025 · similarity 0.85)
- Vision-Language Gradient Descent-driven All-in-One Deep Unfolding Networks (CVPR 2025 · similarity 0.84)
- UHD-processer: Unified UHD Image Restoration with Progressive Frequency Learning and Degradation-aware Prompts (CVPR 2025 · similarity 0.84)
- UARE: A Unified Vision-Language Model for Image Quality Assessment, Restoration, and Enhancement (CVPR 2026 · similarity 0.84)
- DDR: Exploiting Deep Degradation Response as Flexible Image Descriptor (NeurIPS 2024 · similarity 0.84)
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.
