Adaptive enhancement of contrast and brightness of low-light underwater images using feed-forward neural network

Access to high-resolution underwater images is vital in preserving and developing marine resources. Underwater light scattering and light absorption are two Underwater lights scattering, and light absorption are two fundamental issues in improving the quality of underwater images. Many of the captured images have severe degradation that damages the systems and activities that rely on these images. Most of the obtained images have severe degradation that negatively affects the systems and actions based on these images. To address this issue, we have introduced an auxiliary network to enhance the contrast of underwater images. This network comprises three critical components. A decoder network is used to recover the gradient maps and increase the brightness of the images. We use a brightness adjustment network to control the brightness of the hidden image, and finally, we use an adaptive contrast module to adjust the contrast. We use the normalization module to solve the problem of not paying attention to the increase in image contrast when increasing the brightness. The evaluation and comparison of our method using constructed images or images available in public datasets shows that our model effectively increased the resolution of underwater images. In addition, our model can enhance the resolution of complex images in low-light conditions in the deep sea, scenes with dim backgrounds, and images captured in dark environments.