Image Fusion Methods Based On Adaptive Decomposition

Different imaging devices can acquire multiple images of a scene under different optical conditions or at different times.Image fusion fuses the multi-source images of the same scene into a single image with the largest information content without generating information that does not exist in the source image.The fused image describes the scene better than a single source image,and image fusion is an effective method to improve image information quality.Image fusion technology has been widely used in medical imaging,microscopic imaging,remote sensing and computer vision.Image decomposition is a key step in the image fusion method based on transform domain.The degree of image decomposition directly affects the final fusion effect.The multi-scale decomposition method is a widely used image decomposition method.However,the traditional multi-scale decomposition methods often do not decompose according to the nature of the image itself,and it is difficult to distinguish between high frequency and low frequency sub-bands.If the image can be adaptively decomposed according to the nature itself,the information about texture and detail features of the image can be better extracted,which is beneficial to obtain a better fusion effect.In order to solve the above problems,the specific research content of this paper is described as follows:1.Multimodal medical image fusion algorithm based on adaptive decompositionMultimodal medical images reflect different information about human organs and diseased tissues.In clinical diagnosis,single modal images can not provide enough information,while multimodal medical image fusion can make up for the deficiency of single examination imaging.In traditional image fusion algorithms,source images are separated into a fixed space.The low frequency part and the high frequency part are not discriminated according to the nature of the image,resulting in only retaining fuzzy structural information.Traditional fusion rules often use a fixed proportion,causing color distortion.In this paper,a new adaptive decomposition algorithm is proposed to distinguish high frequency and low frequency of structure image to obtain smoothing layer and texture layer.The smoothing layer of the structural image and the color information of the function image are fused according to dynamic rules,and then the texture layer is added.On the basis of the objective evaluation metrics,the spectral information evaluation metrics are introduced to evaluate the retention of color.The experiment results show that the proposed method can retain the color information and structure information very well at the same time.Concerning subjective and objective evaluation,the proposed algorithm is superior to other algorithms.2.Multi-focus image fusion algorithm based on adaptive empirical mode decompositionAffected by the limited field depth of the lens,only part of image with a sharp focus will be obtained in one shot.In order to obtain a fully focused image,it is necessary to adjust the focal length to obtain multiple images and fuse these multi-focus images of the same scene.Most focus detection indicators have to block the image,which is easy to cause blockiness.In addition,if the pixels are directly selected for fusion,the fusion rules are generally linear rules or nonlinear rules.These fusion rules tend to cause the image contrast to decrease,resulting in blurred images.In order to improve the quality of multi-focus image fusion,a multi-focus image fusion method based on adaptive window empirical mode decomposition is proposed in this paper.First,the multi-focus images are decomposed by window empirical mode decomposition,and a set of intrinsic mode function components(high frequency part)and residual components(low frequency part)are obtained.Window empirical mode decomposition can effectively solve the signal aliasing problem in image decomposition.Then,the fusion rule of low-frequency components is determined by the output of the support vector machine,and the clearer focus area is selected.The window gradient contrast algorithm proposed in this paper is used to guide the fusion of high-frequency components,and the consistency of the image is ensured while maintaining the contrast of the fused image.Finally,the window empirical mode decomposition inverse transform is performed to obtain the fused image.Experiments indicates that the proposed method can obtain better fusion quality than other comparison methods in terms of the subjective evaluation and the objective evaluation indicators.3.Under-exposed image enhancement algorithm based on adaptive decomposition and CNNIn the case of poor lighting conditions and the limited dynamic range of imaging device,it is easily to capture the under-exposed images with low contrast and low quality.Traditional single image enhancement methods often fail in revealing image details because of the limited information in a single image.In this paper,a single under-exposed image enhancement based on adaptive synchronous curvelet decomposition and convolutional neural network(CNN)is proposed.The CNN training models need images with different brightness rather than a strict ground-truth image.First,a simple effective synchronous decomposition method is proposed to solve the synergy problem in multi-source image decomposition.Then,two CNN models are designed for high-frequency part and low-frequency part,respectively.They process the high-frequency and low-frequency subbands instead of the source images.The weight map obtained from the CNN model represents the contrast distribution.The exposure map generated by gradient-based visibility assessment indicates the exposure distribution.Finally,the weight map and the exposure map are multiplied to generate the final decision map.Experimental results demonstrate that the proposed method outperforms comparative methods.The main innovations of this paper can be summarized as follows: an adaptive decomposition method for medical images is proposed,which uses dynamic fusion rules to guide multimodal medical image fusion,avoiding edge artifacts and color distortion;based on adaptive window empirical mode decomposition,support vector machine and the proposed window gradient contrast algorithm are used to guide multi-focus image fusion,avoiding the block effect and ensuring the image consistency while maintaining the contrast of the multi-focus image;an adaptive synchronous curvelet decomposition method is proposed to solve the synchronization problem of multi-source image decomposition and an under-exposed image enhancement algorithm combined with CNN is proposed.The advantage of the CNN model is that it does not require a high-quality standard training image which is difficult to obtain.