A hierarchical probabilistic underwater image enhancement model with reinforcement tuning

Wei Song, Zhihao Shen, Minghua Zhang, Yan Wang, Antonio Liotta

Abstract: Highlights•A novel hierarchical probabilistic model with reinforcement learning (HPUIE-RL) is proposed.•A reinforcement learning-based fine-tuning framework is developed for unsupervised optimization.•The HPUIE-RL model demonstrates superior visual and quantitative performance.