😆Awesome-All-in-one-Image-Restoration-Methods

[ Last updated at 26/09/2024 ]

🎯Purpose

We aim to provide a list of all-in-one image/video processing techniques, including restoration, enhancement, coding and quality assessment.

🆕 Always updating...! We update the newest paper once a week and summarize them at the end of each month.

📌 About new papers. If you want to incorporate your studies (e.g., the link of paper or project) on all-in-one image restoration in this repository. Welcome to raise an issue or email us (jiaqima@whu.edu.cn or zhangx0802@whu.edu.cn). We will incorporate it into this repository and our survey report ASAP.

🔖 News!!!

• 2024-xx-xx: A comprehensive survey is coming soon~
• 2024-09-31: Updated new papers before 31/09/2024.
• 2024-08-31: Updated new papers before 31/08/2024.
• 2024-08-28: Our paper Perceive-IR is on arXiv and under peer review, which reach SOTA on two all-in-one settings.
• 2024-07-31: Updated new papers before 31/07/2024.
• 2024-06-21: Updated new papers before 21/06/2024.
• 2024-06-17: This repo is lauched to better investigate existing all-in-one restoration methods!
• 2023-06-23: Our paper ProRes is on arXiv and under peer review, nearly the first to introduce visual prompt into all-in-one image restoration tasks.

🌟 Features

• Benchmark datasets
• Common-used metrics
• Order rank for papers
• All-in-one image restoration
• Adverse weather removal
• All-in-one video restoration
• Other all-in-one related papers

🔠 Abbreviation of degradation

Degradation type	Guassian nosie	Real Nosie	Haze/Fog	Rain	Raindrop	Compressed JPEG
Abbreviation	N	RN	H/F	R	RD	J

Degradation type	Motion blur	Defocus blur	Low-light/Enhancement	Super-resolution	Snow
Abbreviation	B	DB	L/E	SR	S

📝 Table of contents

• All-in-one for Natural Image Restoration
• All-in-one for Adverse Weather Removal
• Application in Adverse Weather Conditions
• All-in-one for Medical Image Restoration
• All-in-one for Document Image Restorationn
• All-in-one for LLM-driven Image Restoration
• All-in-one for Question Answering
• All-in-one for Video Restoration
• Benchmark Datasets
• Common-used Metrics

All in one Image Restoration

Model	Title	Task	Venue	Project	Keywords
DL	A General Decoupled Learning Framework for Parameterized Image Operators	N+R+SR+J	ECCV'18		parameterized image operators
AirNet	All-in-one image restoration for unknown corruption	N+H+R	CVPR'22		Contrastive Learning; Degradation representation
GIQE	Giqe: Generic image quality enhancement via nth order iterative degradation	H+R+B+E+S+RD	CVPR'22		blind distortion combination
IDR	Ingredient-oriented Multi-Degradation Learning for Image Restoration	N+H+R+B+L	CVPR'23		ingredients-oriented; scalable learning
PromptIR	PromptIR: Prompting for All-in-One Blind Image Restoration	N+H+R	NeurIPS'23		prompt-based learning
AMIRNet	All-in-one Multi-degradation Image Restoration Network via Hierarchical Degradation Representation	RN+B+DB+L	ACMMM'23		degradation representation; clustering
ProRes	ProRes: Exploring Degradation-aware Visual Prompt for Universal Image Restoration	RN+R+B+L	arXiv'23.06		Degradation-aware Visual Prompts
DRM-IR	DRM-IR: Task-Adaptive Deep Unfolding Network for All-In-One Image Restoration	H+R+L	arXiv'23.07		task adaptive modeling; model-based image restoring
AutoDIR	AutoDIR: Automatic All-in-One Image Restoration with Latent Diffusion	N+H+R+B+L+RD+SR	arXiv'23.10		Blind IQA; latent diffusion
NDR-Restore	Neural Degradation Representation Learning for All-In-One Image Restoration	N+H+R	arXiv'23.10		neural degradation representation
PIP	Prompt-In-Prompt Learning for Universal Image Restoration	N+H+R/N+H+R+B+L	arXiv'23.12		Degradation-aware and basic prompts
TextPromptIR	Textual Prompt Guided Image Restoration	N+H+R	arXiv'23.12		textual prompt; Blind IR
Restornet	Restornet: An efficient network for multiple degradation image restoration	N+H+R	KBS'24		degradation-specific contrastive loss
CAPTNet	Prompt-based Ingredient-Oriented All-in-One Image Restoration	N+H+R+B	TCSVT'24		visual prompt
TextualDegRemoval	Improving Image Restoration through Removing Degradations in Textual Representations	N+H+R	CVPR'24		image-to-text mapper; T2I
MPerceiver	Multimodal Prompt Perceiver: Empower Adaptiveness Generalizability and Fidelity for All-in-One Image Restoration	N+H+RN+R+RD+B+RB+L+S	CVPR'24		multimodal prompt; diffusion priors
DiffUIR	Selective Hourglass Mapping for Universal Image Restoration Based on Diffusion Model	H+R+B+L+S	CVPR'24		selective hourglass mapping; Diffusion
InstructIR	InstructIR: High-Quality Image Restoration Following Human Instructions	N+H+R/N+H+R+B+L+(SR+E)	ECCV'24		human-written instruction; text-guidance
RAM	Restore Anything with Masks: Leveraging Mask Image Modeling for Blind All-in-One Image Restoration	N+H+R+B+KB+L+J	ECCV'24		Mask Image Modeling
GRIDS	GRIDS: Grouped Multiple-Degradation Restoration with Image Degradation Similarity	11 degradation types	ECCV'24		multiple degradation relationships
UniProcessor	UniProcessor: A Text-induced Unified Low-level Image Processor	30 degradation types	ECCV'24		subject/manipulation prompt
MiOIR	Towards Effective Multiple-in-One Image Restoration: A Sequential and Prompt Learning Strategy	N+H+R+B+L+SR+J	arXiv'24.01		sequential/prompt learning
U-WADN	Unified-Width Adaptive Dynamic Network for All-In-One Image Restoration	N+H+R	arXiv'24.01		Width Adaptive Backbone/Selector
AdaIR	AdaIR: Adaptive All-in-One Image Restoration via Frequency Mining and Modulation	N+H+R/N+H+R+B+L	arXiv'24.03		frequency mining/modulation
DyNet	Dynamic Pre-training: Towards Efficient and Scalable All-in-One Image Restoration	N+H+R	arXiv'24.04		weight-sharing; Dynamic Pre-training
AdaIR	AdaIR: Exploiting Underlying Similarities of Image Restoration Tasks with Adapters	N+R+B	arXiv'24.04		self-supervised pre-training
DaAIR	Efficient Degradation-aware Any Image Restoration	N+H+R/N+H+R+B+L	arXiv'24.05		Degradation-aware Learner; low-rank regime
LM4LV	LM4LV: A Frozen Large Language Model for Low-level Vision Tasks	N+R+B	arXiv'24.05		Frozen Large Language Model
ConStyle v2	ConStyle v2: A Strong Prompter for All-in-One Image Restoration	N+H+R+B+DB+L+S+J	arXiv'24.06		prompter; pre-training
Diff-Restorer	Diff-Restorer: Unleashing Visual Prompts for Diffusion-based Universal Image Restoration	N+RD+B+S+J	arXiv'24.07		VLM; SD
Instruct-IPT	Instruct-IPT: All-in-One Image Processing Transformer via Weight Modulation	N+H+R+B+S	arXiv'24.07		low-rank decomposited
LMDIR	Training-Free Large Model Priors for Multiple-in-One Image Restoration	N+R+L	arXiv'24.07		Large Model Driven
AnyIR	Any Image Restoration with Efficient Automatic Degradation Adaptation	N+H+R/N+H+R+B+L	arXiv'24.07		local-global gated intertwining
MEASNet	Multi-Expert Adaptive Selection: Task-Balancing for All-in-One Image Restoration	N+H+R/N+H+R+B+L	arXiv'24.07		multi-expert adaptive selection
ColorRA	Contribution-based Low-Rank Adaptation with Pre-training Model for Real Image Restoration	N+H+R+RD+B	ECCV'24		low-rank adaptation
HAIR	HAIR: Hypernetworks-based All-in-One Image Restoration	N+H+R/N+H+R+B+L	arXiv'24.08		Classifier/Hyper Selecting
BIR-D	Taming Generative Diffusion for Universal Blind Image Restoration	More than 9 degradations	arXiv'24.08		Blind method; diffusion
UIR-LoRA	UIR-LoRA: Achieving Universal Image Restoration through Multiple Low-Rank Adaptation	More than 10 degradations	arXiv'24.09		LoRA; degradation similarity
LoRA-IR	LoRA-IR: Taming Low-Rank Experts for Efficient All-in-One Image Restoration	N+H+R	arXiv'24.10		MoE; LoRA
COR	Chain-of-Restoration: Multi-Task Image Restoration Models are Zero-Shot Step-by-Step Universal Image Restorers	three degradations	arXiv'24.10		Chain-of-Thought; Zero-shot
MoCE-IR	Complexity Experts are Task-Discriminative Learners for Any Image Restoration	N+H+R/N+H+R+B+L	arXiv'24.11		MoE

Adverse Weather Removal

Model	Title	Task	Venue	Project	Method
All-in-one	All in One Bad Weather Removal using Architectural Search	R+RD+H+S	CVPR'20		NAS; Shared decoder
TKL	Learning Multiple Adverse Weather Removal via Two-stage Knowledge Learning and Multi-contrastive Regularization: Toward a Unified Model		CVPR'22
TransWeather	TransWeather: Transformer-based Restoration of Images Degraded by Adverse Weather Conditions	R+RD+H+S	CVPR'22		Transformer model; Weather type queries
BIDeN	Blind Image Decomposition	R+RD+H+S	ECCV'22		Multi-scale encoder; GANs
UVRNet	Unified Multi-Weather Visibility Restoration		TMM'22
AIRFormer	Frequency-Oriented Efficient Transformer for All-in-One Weather-Degraded Image Restoration		TCSVT'23
WeatherDiffusion	Restoring Vision in Adverse Weather Conditions with Patch-Based Denoising Diffusion Models	R+RD+H+S	TPAMI'23		Patch-based diffusion model
TOENet	Let You See in Haze and Sandstorm: Two-in-One Low-Visibility Enhancement Network		TIM'23
ADMS	All-in-one Image Restoration for Unknown Degradations Using Adaptive Discriminative Filters for Specific Degradations	R+N+B/R+S+H	CVPR'23
SmartAssign	SmartAssign: Learning A Smart Knowledge Assignment Strategy for Deraining and Desnowing	R+S	CVPR'23
WGWS-Net	Learning Weather-General and Weather-Specific Features for Image Restoration Under Multiple Adverse Weather Conditions	R+RD+H/R+S+H	CVPR'23
WeatherStream	WeatherStream: Light Transport Automation of Single Image Deweathering	R+F+S	CVPR'23
RAHC	Learning to Restore Arbitrary Hybrid Adverse Weather Conditions in One Go		arXiv'23.05
WM-MoE	WM-MoE: Weather-aware Multi-scale Mixture-of-Experts for Blind Adverse Weather Removal		arXiv 2023
MetaWeather	MetaWeather: Few-Shot Weather-Degraded Image Restoration	R+D+S+H_M	ECCV'24		meta-learning framework
UtilityIR	Always Clear Days: Degradation Type and Severity Aware All-In-One Adverse Weather Removal	R+RD+H+S	arXiv'23		Marginal Quality Ranking Loss
DDCNet	Decoupling Degradation and Content Processing for Adverse Weather Image Restoration		arXiv'23.12
DwGN	Image All-In-One Adverse Weather Removal via Dynamic Model Weights Generation		KBS'24
DTMWR	Multi-weather Image Restoration via Domain Translation	H+R+S	ICCV'23		domain-translation
AWRCP	Adverse Weather Removal with Codebook Priors		ICCV'23
ViWS-Net	Video Adverse-Weather-Component Suppression Network via Weather Messenger and Adversarial Backpropagation		ICCV'23
GridFormer	GridFormer: Residual Dense Transformer with Grid Structure for Image Restoration in Adverse Weather Conditions	R+RD+H+S	IJCV'24		Grid structure based transformer model
CLIP-SRD	Exploring the Application of Large-scale Pre-trained Models on Adverse Weather Removal		TIP'24
DeformDeweatherNet	Residual Deformable Convolution for better image de-weathering		PR'24
CL_all-in-one	Continual All-in-One Adverse Weather Removal with Knowledge Replay on a Unified Network Structure	H+R+S	TMM'24		continual learning; knowledge distillation
MPDAC	Multiple Adverse Weather Removal Using Masked-Based Pre-Training and Dual-Pooling Adaptive Convolution		Access'24
MoFME	Efficient Deweather Mixture-of-Experts with Uncertainty-aware Feature-wise Linear Modulation	R+RD+S	AAAI'24		MoE architecture
Imperfect-deweathering	Learning Real-World Image De-Weathering with Imperfect Supervision	H+R+S	AAAI'24		Consistent Label Constructor
DiffTTA	Genuine Knowledge from Practice: Diffusion Test-Time Adaptation for Video Adverse Weather Removal		CVPR'24
LDR	Language-driven All-in-one Adverse Weather Removal	R+RD+S	CVPR'24		VLM; Mixture-of-Experts
AoSRNet	AoSRNet: All-in-One Scene Recovery Networks via Multi-knowledge Integration		KBS'24
AiOENet	AiOENet: All-in-One Low-Visibility Enhancement to Improve Visual Perception for Intelligent Marine Vehicles Under Severe Weather Conditions		TIV'24
MWFormer	MWFormer: Multi-Weather Image Restoration Using Degradation-Aware Transformers	R+RD+H+S	TIP'24
JCDM	Joint Conditional Diffusion Model for Image Restoration with Mixed Degradations	R+RD+H+S	arXiv'24.04		Physical model guidance; Diffusion Model
Histoformer	Restoring Images in Adverse Weather Conditions via Histogram Transformer	R+RD+H+S	ECCV'24		Histogram self-attention
OneRestore	OneRestore: A Universal Restoration Framework for Composite Degradation	R+H+S+L	ECCV'24		cross-attention; scene descriptors
WResVLM	Towards Real-World Adverse Weather Image Restoration: Enhancing Clearness and Semantics with Vision-Language Models	R+RD+H+S	ECCV'24		semi-supervised; vision-language model
T3-DiffWeather	Teaching Tailored to Talent: Adverse Weather Restoration via Prompt Pool and Depth-Anything Constraint	R+RD+H+S	ECCV'24		weather-prompts; general prompts; Depth-Anything feature
AWRaCLe	AWRaCLe: All-Weather Image Restoration using Visual In-Context Learning	R+H+S	arXiv'24.09		CLIP; visual context information
AllWeather-Net	AllWeather-Net: Unified Image Enhancement for Autonomous Driving Under Adverse Weather and Low-Light Conditions	R+RD+H+S	arXiv'24.09		knowledge distillation; Scaled Illumination-aware Attention
D4IR	Data-free Distillation with Degradation-prompt Diffusion for Multi-weather Image Restoration	R+H+S	arXiv'24.09		hierarchical architecture; diffusion model
ReviveDiff	ReviveDiff: A Universal Diffusion Model for Restoring Images in Adverse Weather Conditions	R+H+L+UIE	arXiv'24.09		Coarse-to-Fine Learning
WNE-Net	All-in-one Weather-degraded Image Restoration via Adaptive Degradation-aware Self-prompting Model		arXiv'24.11		self-prompting

Application in Adverse Weather Conditions

Model	Title	First Author	Type	Venue	Project
WAS	Semantic Segmentation under Adverse Conditions: A Weather and Nighttime-aware Synthetic Data-based Approach	Abdulrahman Kerim	Image	BMVC 2022
ROD-Weather	Robust Object Detection in Challenging Weather Conditions	Himanshu Gupta	Image	WACV 2024
IDD-AW	IDD-AW: A Benchmark for Safe and Robust Segmentation of Drive Scenes in Unstructured Traffic and Adverse Weather	Furqan Ahmed Shaik	Image	WACV 2024
SD4VS	Leveraging Synthetic Data to Learn Video Stabilization Under Adverse Conditions	Abdulrahman Kerim	Video Stabilization	WACV 2024
CFMW	CFMW: Cross-modality Fusion Mamba for Multispectral Object Detection under Adverse Weather Conditions	Haoyuan Li	Image	arXiv 2024
ControlUDA	ControlUDA: Controllable Diffusion-assisted Unsupervised Domain Adaptation for Cross-Weather Semantic Segmentation	Fengyi Shen	Image	arXiv 2024
DKR	Semantic Segmentation in Multiple Adverse Weather Conditions with Domain Knowledge Retention	Xin Yang	Image	AAAI 2024
PASS	Parsing All Adverse Scenes: Severity-Aware Semantic Segmentation with Mask-Enhanced Cross-Domain Consistency	Fuhao Li	Semantic Segmentation	AAAI 2024
Vehicle-weather	Perception and sensing for autonomous vehicles under adverse weather conditions: A survey	Yuxiao Zhang	Survey	ISPRS JPRS 2023
Fire-Detection	An Effective Attention-based CNN Model for Fire Detection in Adverse Weather Conditions	Hikmat Yar	Fire-Detection	ISPRS JPRS 2023
SDRNet	SDRNet: Saliency-Guided Dynamic Restoration Network for Rain and Haze Removal in Nighttime Images	Wanning Zhu	Image	ICASSP 2024
SeaIceWeather	Deep Learning Strategies for Analysis of Weather-Degraded Optical Sea Ice Images	Nabil Panchi	Natrual Image	IEEE Sensors Journal 2024
PLAIR-AW	Power Line Aerial Image Restoration under Adverse Weather: Datasets and Baselines	Sai Yang	Image	arXiv 2024

Medical Image Restoration

Model	Title	First Author	Type	Venue	Project
AMIR	All-In-One Medical Image Restoration via Task-Adaptive Routing	Zhiwen Yang	Medical Image	MICCAI 2024
ProCT	Universal Incomplete-View CT Reconstruction with Prompted Contextual Transformer	Chenglong Ma	Medical Image	arXiv 2023

Document Image Restoration

Model	Title	First Author	Type	Venue	Project
DocRes	DocRes: A Generalist Model Toward Unifying Document Image Restoration Tasks	Jiaxin Zhang	Document Image	CVPR 2024

Parameter adaptation Image Restoration

Model	Title	Type	Venue	Project	Method
AdaptIR	AdaptIR: Parameter Efficient Multi-task Adaptation for Pre-trained Image Restoration Models	Natrual Image	arXiv'23.12		parameter efficient transfer learning

LLM-driven Image Restoration

Model	Title	First Author	Type	Venue	Project
InstructIR	InstructIR: High-Quality Image Restoration Following Human Instructions	Marcos V. Conde	Natrual Image	ECCV 2024
LM4LV	LM4LV: A Frozen Large Language Model for Low-level Vision Tasks	Boyang Zheng	Natrual Image	arXiv 2024
DACLIP-UIR	Controlling vision-language models for universal image restoration	Boyang Zheng	Natrual Image	arXiv 2024

Question Answering

Model	Title	First Author	Type	Venue	Project
PromptGIP	Unifying Image Processing as Visual Prompting Question Answering	Yihao Liu	Natrual Image	arXiv 2023

Video Restoration

Model	Title	First Author	Type	Venue	Project
CDUN	Degradation Conditions Guided Cross-Consistent Deep Unfolding Network for All-In-One Video Restoration	Yuanshuo Cheng	AIO	arXiv 2023
DiffTTA	Genuine Knowledge from Practice: Diffusion Test-Time Adaptation for Video Adverse Weather Removal	Yijun Yang	Adverse weather	CVPR'24
SD4VS	Leveraging Synthetic Data to Learn Video Stabilization Under Adverse Conditions	Abdulrahman Kerim	Video stabilization	WACV 2024
ViWS-Net	Video Adverse-Weather-Component Suppression Network via Weather Messenger and Adversarial Backpropagation	Yijun Yang	Adverse weather	ICCV'23

Degradation Forgetting

Model	Title	First Author	Type	Venue	Project
OD	Accurate Forgetting for All-in-One Image Restoration Model	Xin Su	Degradation Forgetting	arXiv 2024

Benchmark Datasets (Multi-Weather)

Dataset	Task	Usage	Link	Year
Raindrop	raindrop	Train: 861/Test: set-A (58), set-B (239)	download	CVPR 2018
Snow100K	snow	100K synthetic images	download	IEEE TIP 2018
Outdoor-Rain	haze & rain	9000 train/1500 val	download	CVPR 2019
ALL-Weather	haze & rain/snow/raindrop	Train (18069): Outdoor-Rain (8250), Snow100K (9001), and Raindrop (818)/Test: Snow100K-L (16801), RaindropA (58), Outdoor-Rain (750)	download	CVPR 2022
Flickr2K	Image Super-resolution	Training		2017
SeaIceWeather	Image De-weathering	Training,Testing		2024

Benchmark Datasets (Distortion) Working In Progress!

Dataset	Task	Usage	Year
DIV2K	Image Super-resolution	Training,Testing	2017
Flickr2K	Image Super-resolution	Training	2017
SeaIceWeather	Image De-weathering	Training,Testing	2024

Common-used Metrics Working In Progress!

🤝 Cite us

WIP!