Research On High Dynamic Range Image Tone Mapping

Dynamic range is the logarithm of the ratio of the maximum to the minimum luminance for a digital image.In reality,the real scene with wide dynamic range can be effectively perceived by human eyes and displayer,resulting in some loss of scene information.To address this deficiency,high dynamic range(HDR)imaging technology employs floating-point numbers to represent a wide range of luminance,which can retain the fidelity of real scene accurately.But unfortunately,these HDR display devices are difficult to popularize because of their cost and technical problems.Hence,how to visualize HDR images effectively on existing display devices has become a problem that must be solved in practical applications.To tackle the problem,tone mapping(TM),which essentially functions as intensities mapping of an HDR image to the target display range,has been developed.This paper first studies some of the existing tone mapping operators,analyzes their advantages and disadvantages,such as detail lost,local brightness imbalance,halo and so on.In this paper,a novel TM method based on macro-micro modeling is proposed,which can address the common problems in existing TM methods,such as exposure imbalance and halo artifact.From a microscopic perspective,multi-layer decomposition and reconstruction are applied to model the properties of brightness,structure,and detail for HDR images,and then different strategies are adopted for each layer by the human visual system(HVS)to reduce the overall brightness contrast and retain as much scene information.From a macroscopic perspective,scene content-based global operator is designed to adaptively adjust the scene brightness so that it is consistent with the subjective perception of human eyes.Both the micro and macro models are processed in parallel,which can ensure the integrity and subjective consistency of scene information.Experiments with numerous HDR images and TM methods are conducted and the results show that the proposed method achieves visually compelling results with little exposure imbalance and halo artifact,and is superior to the current state-of-the-art TM methods in both subjective and objective evaluations.In order to further improve the effect of the operator,a tone mapping operator with fusion significance detection is proposed.When viewing an image,the human visual system spontaneously focuses its attention on the most important areas of the image.By integrating the significance detection method based on deep learning into the tone mapping,we can better process the significance area,reserve as much image information of the significance area as possible,and improve the visual perception.First by extending the human eye perception,respectively from the angle of the micro and macro visual cognitive mechanism modeling,and through the study of the parallel processing of the two models to obtain the required LDR images,then significant detection method are introduced,marked significant area,finally through the process again of significant area,optimization of mapping results.Experimental results show that the algorithm improves visual perception,has a wider range of adaptability,has a strong sense of colorful layers,and can well retain more details.