A visual hierarchical framework based model for underwater image enhancement

Abstract

Although there are various computer vision based processing methods for underwater image enhancement, some of them have to be implemented under given priors and the performance of these approaches is still limited. We propose a visual hierarchical framework based model inspired by retinal mechanism, which restores the distortion from color cast and increases contrast simultaneously for underwater images by simulating visual processing of the mammalian retina. The RGB components and brightness of the input image are transmitted from the sampling of rod and cone photoreceptors to the higher visual cortex V4 via sub-layers of bipolar cells, amacrine cells, retinal ganglion cells (RGCs), lateral geniculate nucleus (LGN) and V1 respectively. The flowchart of our presented framework is shown in Fig. 1. Each layer in this procedure is designed in the light of color processing mechanism existing in the human visual system (HVS) from the retina to the higher visual cortex [1,2]. Then, a robust image formation model is employed for scene recovery with restricted transmission map and estimated background light. The main contributions of this work can be summarized as follows: 1) We propose a more integrated retinal mechanism inspired model for color correction and contrast improvement of underwater images.