# LIQA: Lifelong Blind Image Quality Assessment

arXiv: https://arxiv.org/abs/2104.14115 · HTML: https://mlsanity.com/paper/2104.14115/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Apr 2021.

## Built on by (most cited first, 20 of 20)

- [Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective](https://mlsanity.com/paper/2303.14968/) (2023)
- [NTIRE 2024 Challenge on Short-form UGC Video Quality Assessment: Methods and Results](https://mlsanity.com/paper/2404.11313/) (2024)
- [Task-Specific Normalization for Continual Learning of Blind Image Quality Models](https://mlsanity.com/paper/2107.13429/) (2021)
- [SwinIQA: Learned Swin Distance for Compressed Image Quality Assessment](https://mlsanity.com/paper/2205.04264/) (2022)
- [Source-free Unsupervised Domain Adaptation for Blind Image Quality Assessment](https://mlsanity.com/paper/2207.08124/) (2022)
- [StyleAM: Perception-Oriented Unsupervised Domain Adaption for Non-reference Image Quality Assessment](https://mlsanity.com/paper/2207.14489/) (2022)
- [RTN: Reinforced Transformer Network for Coronary CT Angiography Vessel-level Image Quality Assessment](https://mlsanity.com/paper/2207.06177/) (2022)
- [MixNet: Efficient Global Modeling for Ultra-High-Definition Image Restoration](https://mlsanity.com/paper/2401.10666/) (2024)
- [Forgetting to Remember: A Scalable Incremental Learning Framework for Cross-Task Blind Image Quality Assessment](https://mlsanity.com/paper/2209.07126/) (2022)
- [DGIQA: Depth-guided Feature Attention and Refinement for Generalizable Image Quality Assessment](https://mlsanity.com/paper/2505.24002/) (2025)
- [Latent Guidance in Diffusion Models for Perceptual Evaluations](https://mlsanity.com/paper/2506.00327/) (2025)
- [Perception-oriented Bidirectional Attention Network for Image Super-resolution Quality Assessment](https://mlsanity.com/paper/2509.06442/) (2025)
- [Quality Assessment of Low Light Restored Images: A Subjective Study and an Unsupervised Model](https://mlsanity.com/paper/2202.02277/) (2022)
- [Subjective and Objective Quality-of-Experience Evaluation Study for Live Video Streaming](https://mlsanity.com/paper/2409.17596/) (2024)
- [RulerNet: Learning Perspective-Invariant Ruler Representations for Robust Image Scale Estimation](https://mlsanity.com/paper/2507.07077/) (2025)
- [Surveillance Facial Image Quality Assessment: A Multi-dimensional Dataset and Lightweight Model](https://mlsanity.com/paper/2602.07403/) (2026)
- [Deep Tiny Network for Recognition-Oriented Face Image Quality Assessment](https://mlsanity.com/paper/2106.04852/) (2021)
- [Evolving Knowledge Mining for Class Incremental Segmentation](https://mlsanity.com/paper/2306.02027/) (2023)
- [Preference-Guided Debiasing for No-Reference Enhancement Image Quality Assessment](https://mlsanity.com/paper/2603.20086/) (2026)
- [Blind Stereoscopic Omnidirectional Image Quality Assessment Using Predictive Coding Hierarchy](https://mlsanity.com/paper/2608.28798/) (2026)

## Abstract (arXiv)

Existing blind image quality assessment (BIQA) methods are mostly designed in a disposable way and cannot evolve with unseen distortions adaptively, which greatly limits the deployment and application of BIQA models in real-world scenarios. To address this problem, we propose a novel Lifelong blind Image Quality Assessment (LIQA) approach, targeting to achieve the lifelong learning of BIQA. Without accessing to previous training data, our proposed LIQA can not only learn new distortions, but also mitigate the catastrophic forgetting of seen distortions. Specifically, we adopt the Split-and-Merge distillation strategy to train a single-head network that makes task-agnostic predictions. In the split stage, we first employ a distortion-specific generator to obtain the pseudo features of each seen distortion. Then, we use an auxiliary multi-head regression network to generate the predicted quality of each seen distortion. In the merge stage, we replay the pseudo features paired with pseudo labels to distill the knowledge of multiple heads, which can build the final regressed single head. Experimental results demonstrate that the proposed LIQA method can handle the continuous shifts of different distortion types and even datasets. More importantly, our LIQA model can achieve stable performance even if the task sequence is long.