Transfer Function Research Based On Stable Field Reconstruction

As the reconstructed image(or video data)has become a major part of the big data and information flow,the acquisition,transmission,preservation,or human factors may cause the defect of the local key information in the image.However,the existing image reconstruction(or repair)algorithms just pursuit human visual effects while ignoring the accuracy of image reconstruction.At the same time,the efficiency of these algorithms are really low because it needs several times of iteration calculation,thus,this process is bad for processing information in real time.At present,there are several problems of above image repair and reconstruction.So,the researchers have proposed local image reconstruction algorithms that based on the stable field.These algorithms effectively combined the image texture and the stable field together by their relationship.When we reconstruct the defect area,the known information that outside the defect area is transmitted into the inside area according to the boundary of the defect area,and the transfer process only involves one iteration process.There is an algorithm set the point-source influence function as the transfer function to calculate the information transmission,and this algorithm achieves preliminary results in accurate reconstruction.Due to the problem of the accuracy and efficiency of two-dimensional image reconstruction(or inpainting),this paper designed the transfer function as the core function and proposed a new relative reconstruction algorithm.The stable field model is used to describe the image local region in this paper.In view of each pixel in the defect region,a reconstruction model considers the pixels in the known region are transmitting energy to each known pixel.During the reconstruction process,the energy is firstly transmitted into the nearest neighborhood,then,we introduce the Second-order Taylor expansion,and the reconstruction progress is achieved by constructing the transfer function which using the Second-order Taylor expansion.Finally,according to the energy value in the nearest neighbor domain,the reconstruction is completed by interpolation.Based on the foundation of image local regions stable field model and the inpainting algorithms that based on the stable field,an improved algorithm is proposed in this paper which studies the transfer function of two-dimensional image reconstruction and the related reconstruction algorithms,and it points at the reconstruction of image edge and texture details.Under the premise of keeping a good visual effect,the proposed method has greatly improved the reconstruction accuracy and PSNR(Peak Signal to Noise Ratio),especially in the image defect region edge and texture details of the reconstruction performance outstanding.In addition,the experimental results show that the consequence of thispaper's algorithm are especially good,and this algorithm has universal applicability to different kind of images of different degrees of defect.The current image spot processing or removal is aimed at difficult for human to aware,and aim at getting good visual effect rather than the correct rate.In this paper,based on the transfer function,a spot area reconstruction method that based on the stable field is proposed.The transfer function thus constructed through the basis that the light spot is generally formed by the superimposition of another local strong light source during the image imaging.Then we have to calculate energy that in the nearest neighbor outside the light spots area.According to the image reconstruction model based on the stable field,the energy transfer function which transfer energy from the known pixel area to spot pixel area can be calculated,and the local strong light source energy of the spot region is removed by the exchange of point and source,and then the reconstruction process is completed.The experimental results show that the spot reconstruction algorithm can reconstruct the surface spot of any kind of images and the reconstruction effect is better than other spot removal algorithms.In summary,this paper presents the relevant reconstruction algorithms that based on the stable field image reconstruction model,and used the transfer function as the core function.The experimental results show that our algorithms can reconstruct the local missing texture and the texture of the image with spot areas,also,it can greatly improve the PSNR and the accuracy of the reconstructed image,while being used to process information in real time,maintaining a good visual effect and reconstruction efficiency.