Underwater Image Enhancement Based on Hybrid Enhanced Generative Adversarial Network

In recent years, underwater image processing has played an essential role in ocean exploration. The complexity of seawater leads to the phenomena light absorption and scattering, which turn cause serious degradation problems, making it difficult capture high-quality images. A novel enhancement model based on Hybrid Enhanced Generative Adversarial Network (HEGAN) is proposed this paper. By designing a Underwater Image Synthesis Model (HUISM) physical deep learning method, many richly varied paired images are acquired compensate for missing problem dataset training. Meanwhile, Detection Perception Enhancement (DPEM) with Perceptual Loss designed transfer coding knowledge form gradient through perceptual loss, generation visually better detection-friendly Then, synthesized enhanced models integrated into adversarial network generate clear game learning. Experiments show that method significantly outperforms several state-of-the-art methods both qualitatively quantitatively. Furthermore, also demonstrated can improve target detection performance environments, specific application value subsequent processing.