Exposure Time Simulation Thanks to the LIP (Logarithmic Image Processing) Model with Noise Reduction by Deep Convolutional Neural Networks

The LIP (Logarithmic Image Processing) model is recognized as an efficient framework to process images acquired in transmitted and reflected light, and to take into account the human visual system. An important property of the LIP model consists of simulating exposure time variations. Applied to very low light images, our LIP algorithms enhance not only the signal, but also the noise and lead to quantized grey levels. In order to overcome such a drawback, we perform a noise reduction based on deep convolutional neural networks.