Browsing by Subject "Digital image reconstruction"
  • Item
    An Adaptive And Statistical Efficiency Myriad Filter For A Recursive Image Reconstruction Using a Multi-Frame SRR Algorithm With A Stochastic Regularization For Video Sequences
    ( 2015) Vorapoj Patanavijit
    In real applied implementations, a collection of classical linear filtering theories, for example, a Median filter and a Mean filter, can be only applied to the Gaussian noise environments due to the fact that these linear filters usually gives the poor performance under the presence of non-Gaussian noise environments. Because of the motion estimation blunder and observation process error, which are usually caused from real electronic noise, non-accurate optical devices or mathematical simplification models of observed process systems, the SRR (Super Resolution Reconstruction) algorithms using classical linear filters (a Median filter and a Mean filter) should possibly demote the quality of a reconstructed image rather than improve its quality. Under non-Gaussian environments, a class of flexible filters with high statistical efficiency, so called Myriad filter, has been proposed and its solid theoretical principle was analyzed for indicating that the Myriad filter is usually more powerful than a class of linear filters, especially for non-Gaussian environments. This paper proposes an adaptive and statistical efficiency myriad filter for a recursive image reconstruction for applied implementing on real video sequences, which are contaminated by both Gaussian and non-Gaussian noise environments. Thus, the Myriad filter, which is implemented for getting rid of noise in an expected image and for valuing the contrast between the back-propagated expecting of the reconstructed high resolution image and a group of low resolution images, is involved in this stochastic regularization SRR framework for the elimination proposing of Gaussian and non-Gaussian noise. Because of an ill-pose condition of the SRR algorithm, Tikhonov regularization methodology is mathematically required for getting rid of deformation from the reconstructed high resolution image and reforming the calculated time of its convergence. Under a lot of noisy corrupted environments (Noiseless, AWGN, Impulsive Noise, Poisson Noise and Speckle Noise at unequal noise power), the performance from the both PSNR and virtual quality prospect of the proposed SRR algorithm using Myriad filter, which is compared with SRR algorithms using classical linear filters (a Median filter and a Mean filter) are depicted and the proposed SRR algorithm gives the superior visually quality and, thus, superior PSNR than the SRR algorithms using classical linear filters.
  • Item
    An Alternative Single-Image Super Resolution Framework Employing High Frequency Prediction Using A Robust Huber Rational Function
    ( 2015-11) Kornkamol Thakulsukannant ; Vorapoj Patanavijit ; Vincent Mary School of Engineering
    In general prospective, SI-SR or Single-Image Super-Resolution, which is one of the most useful algorithms of Super Resolution-Reconstruction (SRR) algorithms, is a mathematical procedure for acquiring a high-resolution image from only one coarse-resolution image, which is usually computed by Digital Image Processing (DIP). Even thought there have been substantially researched during the last decade, Single - Image Super-Resolution for applying on real implementations still keeps throw down the gauntlet. One of the practical Single- Image Super-Resolution is the resolution enhancement using prediction of the high-frequency image because of its high performance and its less comple xity however the rational function C(x, y) of high-frequency image prediction process of this technique is depend upon three parameters (b, h, k) therefore the parameter turning is difficult for maximizing its performance. From this problem prospective, this paper presents the alternative SI-SR framework employing robust rational function based on Huber function, which is depend upon only one parameter (T), instead of three parameters like the rational function C(x,y). Using up to fourteen standard images, which are crooked by varied noise models, in analysis testing section, the proposed SI-SR is demonstrated to be somewhat simper than the original SI-SR with equivalent efficiency because the saving in parameter turning time will be very important for SI-SR in real implementations.
  • Item
    An Experimental Performance Analysis of Image Reconstruction Techniques under Both Gaussian and Non-Gaussian Noise Models
    ( 2012-07) Vorapoj Patanavijit
    Recently, the images reconstruction approaches are very essential in digital image processing (DIP), especially in terms of removing the noise contaminations and recovering the content of images. Each image reconstruction approach has different mathematical models. Therefore a performance of individual reconstruction approach is varied depending on several factors such as image characteristic, reconstruction mathematical model, noise model and noise intensity. Thus, this paper presents comprehensive experiments based on the comparisons of various reconstruction approaches under Gaussian and non-Gaussian noise models. The employing reconstruction approaches in this experiment are Inverse Filter, Wiener Filter, Regularized approach, Lucy-Richardson (L-R) approach, and Bayesian approach applied on mean, median, myriad, meridian filters together with several regularization techniques (such as non-regularization, Laplacian regularized, Markov Random Field (MRF) regularization, and one-side Bi- Total Variation (OS-BTV) regularization). Three standard images of Lena, Resolution Chart, and Susie (40th) are used for testing in this experiment. Noise models of Additive White Gaussian Noise (AWGN), Poisson, Salt&Pepper, and Speckle of various intensities are used to contaminate all these images. The comparison is done by varying the parameters of each approach until the best peak-signal-to-noise ratio (PSNR) is obtained. Therefore, PSNR plays a vital parameter for comparisons all the results of individual approaches.
  • Item
    A Novel Frequency Domain Image Reconstruction Based on the Tikhonov Regularization and Robust Estimation Technique for Compressive Sensing
    ( 2013-05) Vorapoj Patanavijit
    Recently, a lot of attention has been paid to image reconstructionalgorithms based on Smoothed L0 (SL0) under the frequency domain. SL0 is fast and accurate under the noise free environment however it is unstable with the additional noise.According to ill-posed condition; without any prior information of the original image, the reconstruction procedure of SL0 is much affected by the noise. The frequency domain Tikhonov reduces and constrains the gap of restored image due to the ill-posed situation. Therefore, image restoration algorithm is better and immutable under the noise which can eliminate the image’s properties. Moreover, in this paper we propose an l1 estimation, it is conceived less sensitivity to the outlier than an l2. Thereforethe quality of reconstructed image under noise with high power is improved. Furthermore, the advancedrobust regularization algorithmcan be effectively applied under difference type of noise models (such as Speckle noise,AWGN, Salt & Pepper noise and Poisson noise) and at different noise powers.
  • Item
    A Performance Impact of An Edge Kernel for The High-Frequency Image Prediction Reconstruction
    ( 2014) Vorapoj Patanavijit ; Chaiyod Pirak ; Ascheid, Gerd
    As a rule, the performance of almost digital image processing (DIP) algorithms and these applications directly depends on the spatial resolution of observed input images. Unfortunately, from the current image sensor technology, it is hard to take sufficient high spatial resolution images from commercial devices therefore the fantastic research attempts and, consequently, simple digital image resolution enhancements have been boosted in the last decade. The high-frequency image prediction reconstruction is the simple and effective algorithm for enhancing the image resolution however this algorithm is strongly depends on the edge detection kernel and M0 parameter. Therefore, this paper studies a performance impact of an edge detection kernel such as Roberts kernel, Prewitt Kernel, Sobel Kernel, Laplacian Kernel and Laplacian of Gaussian (LOG) Kernel for the high-frequency image prediction reconstruction. This paper presents three experimental performance studies under a noiseless environment, several blurred environments at different blurred variance and several noisy environments at different noise power levels. The first performance study is an empirical exhaustive study of an optimal edge detection kernel and the study of optimal M0 value is experimentally determined under this environment. The second performance study is an empirical exhaustive study of an optimal edge detection kernel and the study of optimal M0 value is experimentally determined under these environments. Finally, the last performance study is an empirical exhaustive study of an optimal edge detection kernel and the study of optimal M0 value is experimentally determined under these environments.
  • Item
    Tutorial on image reconstruction based on weighted sum (WS) filter approach: from single image to multi-frame image
    (Assumption University Press, 2009) Vorapoj Patanavijit ; Assumption University. Vincent Mary School of Science and Technology
    For large magnification factors, the prior classical smoothness leads to overly smooth results with very little high-frequency content. The classical image restorations are failing to reconstruct the desired image. Consequently, the Recognition-Based Restoration is desired for these purposes and one of the most effective techniques is of a weighted sum (WS) filter. This paper reviews the research framework of weighted sum (WS) filter approach for image reconstruction. This research framework first starts with the Hard-Partition-based Weighted Sum (HP-WS) filter proposed in 1999 and then consequently reviews the Subspace HP-WS (S-HPWS) using PCA (Principal Component Analysis) Filter proposed in 2005, The Soft-Partition-based Weighted Sum (SP-WS) proposed in 2006, and the fast Adaptive Wiener Filter proposed in 2007. The paper reviews each filter technique in terms of its computational concepts, demonstrates parameter optimization from the point of view of the mathematical analysis, and discusses advantages and disadvantages.