Papers by Keyword: Image Restoration

Paper TitlePage

Authors: Wai Bin Huang, Wei Ran Lin
Abstract: This paper presents a level set method for image restoration. In the light of the loss of image quality caused by occlusion or part stain, it adopts the method of picking-up image level sets and filling in level sets of occlusion to reconstruct the image. In the process of linking level lines, besides the traditional geodesic curves, it makes use of the Meaningful beeline detection technique. The experiment results show this method is of great importance in the preservation of images.
Authors: Shu Zhen Wang, Zi Jian Zou, Li Li, Xiao Ping Zhang
Abstract: Blind deconvolution is the restoration of original image from a blurred one when the blur kernel is unknown. While recent algorithms have afforded dramatic progress, the results are still far from perfect in terms of efficiency and stability. In order to gain a stable, unique and effective solution, this paper uses a scale invariant and sparse regularization function to exert regularization constraints on the original image and PSF simultaneously. An experiment is conducted to verify that our image blind recovery algorithm is robust and has stable convergence.
Authors: Qian Qian Quan
Abstract: To the deficiencies of traditional methods for avoiding motion image blurring, a motion blur image restoration method is studied based on Wiener filtering in this paper. The formation factors of motion-blurred images and the imaging process are analyzed, and the motion blur degradation model is established. It introduced the working principle of Wiener filtering, described the steps of blurred image restoration in details. The experiment testing and data analyzing are also conducted. Experimental results showed that the method can has good performance.
Authors: Ming Wei Sheng, Yong Jie Pang, Hai Huang, Tie Dong Zhang
Abstract: The main purpose of underwater image fusion is to combine multi-images about the same object into a high-quality image with abundant information. A new underwater image fusion scheme based on Biorthogonal wavelet transform was presented, which is suitable to underwater computer vision system of AUV. Firstly, median filter algorithm was involved for improving the quality and contrast of two source underwater blurred images. Secondly, the different-position-focused underwater images were decomposed by Biorthogonal wavelet and the wavelet coefficients were acquired for reconstructing the fusion image. Finally, the fused image was constructed using the low-frequency and high-frequency domain fusion rules. By adopting a series of experiments for the underwater images fusion, an integrated underwater image with visible outline and distinguishable inner details was obtained.
Authors: Min Ma, Liang Zhao
Abstract: Image restoration is necessary in many applications as the captured images are inevitably noise-contaminated. Typically, the partical differential equations based methods, which is a primary class of image inpainting techniques, is well accepted. In this paper,anisotropic diffusion (P-M) model was introduced to image denoisng. Simulation results were implemented of the proposed method by using Matlab, in which different levels of noise were compared to show the advantages and the disadvantages.
Authors: Yu Fan, Xue Feng Wu
Abstract: Computational photography and image processing technology are used to restore the clearness of images taken in fog scenes autmatically.The technology is used to restore the clearness of the fog scene,which includes digital image processing and the physical model of atmospheric scattering.An algorithm is designed to restore the clearness of the fog scene under the assumption of the albedo images and then the resolution algorithm is analysised.The algorithm is implemented by the software of image process ,which can improve the efficiency of the algorithm and interface.The fog image and defogging image are compared, and the results show that the visibility of the image is improved, and the image restoration is more clearly .
Authors: Feng Wang, Kun Fan Zhang, Fan Kun Meng, Yong Jun Zhao
Abstract: The RL(Richardson-Lucy) algorithm is an important method for restoration of turbulence-degraded images. However, the shortcoming of this method is that it tends to amplify the noise and exsits excessive smoothing in the iterative procedure. This paper discusses the RL algorithm and its improving methods focusing on turbulence-degraded images restoration.Firstly, a short exposure atmospheric turbulence-degraded model is established and a numerical computing method is proposed for random phase screen. Secondly, the essential principle and computational formula are deduced. To restore the object image effectively from the turbulence-degraded image, a new double-circulation iterative Richardson-Lucy restoration algorithm using TV-regularized method is proposed. This new algorithm introduces the total variation restraint and estimates the object image and the point spread function based on the inner and outer double-circulation iteration, which can use the inherent relation between the object image and the point spread function adequately. Simulation experiments show that the proposed algorithm can effectively preserve the details and edges of the image and its restoration effect is obviously better than the traditional RL algorithm.
Authors: Li Feng Zhang, Hui Min Lu, Yuhki Kitazono, Serikawa Seiichi
Abstract: Nowadays, digital imaging device has become a popular item of our daily life. Even a cell phone can gives a high density image. People can enjoy taking a photo at anytime and anywhere, but such a mechanically simplified imaging device usually cannot mount an optical lens filter. Therefore the digital filtering effect is needed. Many image effect filters have been released by image processing soft maker or digital camera manufacture, but most of them were made just for funny, the resulting image is unnatural. This research focuses on the digital filtering effect, and it is realized according to a real optical lens imaging system. In this work, we prepared several optical filter lenses, and take photos with or without using them. Assuming the photo without using a filter is the real object, and then the photo using the filter is the observed filtering image, the digital filter can be derived by a division theoretically. Using these digital filters, the resulting filtering images are gotten. From these result, we find that some of the filtering effect is good, but some of the effects were not enough. Therefore, assuming the cause of these problems from the experimental results, show the direction for future research such as improving experiment approach or adopting nonlinear algorithms is also attempt.
Authors: Min Cao
Abstract: The alphabet computer recognition methods are widely applied in many areas. The clarity of the alphabet is first key step. This paper proposes a kind of image clarity algorithm for alphabet. The key frames in the surveillance video are analyzed in frequency domain to identify the key parameters causing alphabet illegibility and restore the surveillance video. The experiment results illustrate the algorithm can well restore the key frames of the alphabet in the video which can be widely applied in the vehicle plate recognition.
Authors: Wei Sun, Sheng Nan Liu
Abstract: An adaptive variational partial differential equation (PDE) based aproach for restoration of gray level images degraded by a known shift-invariant blur function and additive noise is presented. The restoration problem of a degraded image is solved by minimizing this model, and this minimizing problem is realized by using Hopfield neural network. In the proposed image restoration model, an adaptive regularization parameter is developed instead of the constant regularization parameter used in previous PDE model. The value of the adaptive regularization parameter changes according to different regions of the image to remove noises and preserve edge better. Several computer simulation results show that the image restoration results of the proposed model both look better and have better SNR (Signal to Noise Ratio) than the previous variational PDE based model.
Showing 1 to 10 of 47 Paper Titles