The field of image restoration and enhancement is moving towards more sophisticated and generalizable methods, with a focus on leveraging diverse data and robust priors to improve performance in real-world scenarios. Recent developments have seen the introduction of novel datasets and frameworks that address the challenges of reflection removal, image dehazing, and polarization-based imaging. These advancements have led to significant improvements in image clarity and fidelity, particularly in complex and diverse environments. Notable papers in this area include Dereflection Any Image, which proposes a comprehensive solution for robust reflection removal, and ProDehaze, which employs internal image priors to direct external priors for faithful image dehazing. Other noteworthy papers include Real-World Remote Sensing Image Dehazing, which introduces a large-scale dataset and novel framework for real-world RSID, and PolarFree, which leverages polarization cues for accurate reflection removal. Additionally, Learning Hazing to Dehazing presents a novel hazing-dehazing pipeline that harnesses generative diffusion priors for realistic haze generation and diffusion-based dehazing.