The field of image restoration and quality assessment is moving towards more advanced and innovative methods. Recent developments have focused on improving the accuracy and efficiency of image restoration models, particularly in the context of blind image restoration and spectral compressive imaging. The integration of deep unfolding networks with latent diffusion models and low-rank decomposition has shown promising results in achieving state-of-the-art performance. Additionally, the development of new imaging models and frameworks, such as the Low-Rank Deep Unfolding Network, has mitigated the ill-posedness of mapping 2D residuals back to 3D space of high-dimensional images. Furthermore, research has also explored the use of graph convolutional networks and multi-expert-based feature decoupling to enhance the quality of blind image quality assessment. Noteworthy papers in this area include UnfoldLDM, which proposes a deep unfolding-based blind image restoration method with latent diffusion priors, and Life-IQA, which introduces a GCN-enhanced layer interaction and MoE-based feature decoupling framework for blind image quality assessment. The Test-Time Preference Optimization paradigm has also been proposed to enhance perceptual quality and adapt flexibly to various image restoration tasks without requiring model retraining.