Browsing by Author "Vorapoj Patanavijit"
Results Per Page
Sort Options
-
Item4x4 High-Magnification Image Reconstruction Based on Hybrid of Multi-frame SR Approach and Image Super Resolve Algorithm( 2015-11) Vorapoj PatanavijitFrom tremendously soliciting high quality and high resolution images, sundry image reconstruction algorithms have been researched and implemented during the last fifteen years, especially for high-magnification image reconstructions. In this paper, we develop high-magnification image reconstruction based on hybrid of a multi-frame SR approach and an image super resolve algorithm for 4x4 magnifying in resolution. First, a group of polluted low resolution images are amalgamated for 2x2 magnifying in resolution and suppressing the noise in each polluted low resolution images. Next, the 2x2 magnified image is reconstructed by using image super resolve algorithm based on the high-frequency image prediction to be the 4x4 magnified image. In the performance testing section, the outcomes on both benchmark (in PSNR) and virtual quality, contrasting with previous classical algorithms from the research literature (such as a classical interpolation technique, a classical SRR and an image super resolve algorithm), expose that the proposed hybrid framework has the better performance in both benchmark (in PSNR) and virtual quality.
-
ItemAn 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 PatanavijitIn 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.
-
ItemAn Alternative Single-Image Super Resolution Framework Employing High Frequency Prediction Using A Robust Huber Rational FunctionIn 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.
-
ItemAn Alternative SR Spatial Enhancement Based on Adaptive Meridian Filter and GOM Registration for Severe Noisy Blurred Videos( 2015-08) Kornkamol Thakulsukanant ; Vorapoj PatanavijitCommonly, filtering technique and the video registration technique are two main significance factors of a video SR (Super Resolution) enhancement algorithm. First, the classical filtering technique is based on a linear filter such as mean or median filter that are only suitable for noiseless or low power noise. Later, classical video registration techniques are usually based on a simple translation model because of the fast computation and easy implementation thereby this registration has high precision error. To get over both problems, this paper proposed the alternative SR spatial enhancement using adaptive meridian filter and GOM (General Observation Model) registration for severe noisy blurred videos. The adaptive meridian filter is a robust filter, which is desire for controlling high power outlier, and GOM is a high precision registration technique, which is desired for registering a fast spatial sequence. For proving the proposed performance, the simulated experiments are done in several environments as following: 1. Additive White Gaussian Noise (AWGN) at SNR=15, 17.5, 20, 22.5, 25dB; 2. Poisson Noise; 3. Multiplicative White Gaussian Noise (Speckle Noise) at V=0.01, 0.02, 0.03; 4. Salt and Pepper Noise at D=0.005, 0.010, 0.015. The proposed enhancement algorithm shows that the PSRN of the enhanced image is higher than the SR spatial enhancement based on classical filter with classical registration and GOM registration.
-
ItemAn Alternative Technique using Median Filter for Image Reconstruction based on Partition Weighted Sum Filter( 2016-06) Vorapoj PatanavijitIn this paper, we propose an alternative technique for image reconstruction which it is combined existing methods for better performance in spatial domain using the median (MED) filter based on partition weighted sum (PWS) filter. Four noise models are considered including additive white Gaussian noise (AWGN), poission noise (PN), salt and pepper noise (SPN) and speckle noise (SN) under different image types such as aerial image, face image, scenic image, and text image. The simulation results show that the median based partition weighted sum (MPWS) filter provides better results than the MED and PWS filters in case of AWGN and SPN when the noise probability is not less than 20% for all image types. However, this filter takes longer average simulation time than the MED and PWS filters.
-
ItemBidirectional Confidential with Bilateral Filter on Local Based Optical Flow for Image Reconstruction under Noisy Condition( 2015) Darun Kesrarat ; Vorapoj PatanavijitMore than a decade, Optical flow is relevant in many areas such as video coding and compression, robot vision, object tracking and segmentation, and super resolution reconstruction. By the result of the motion vector from optical flow, reduction the error stands a problem especially under noisy condition. Many models have been proposed to reduce the error and bilateral is one of the popular models. This paper introduces the model of bilateral filter in combination with bidirectional confidential over simple local based optical flow where the quality of the restored images from the returned motion vector is focal point. Several noise levels over several official sequences are observed for the noise tolerance analysis where PSNR is used as an indicator.
-
ItemThe Bilateral Denoising Performance Influence of Window, Spatial and Radiometric Variance( 2015-08) Vorapoj PatanavijitIn the research operation of Digital Signal Processing (DSP) and Digital Image Processing (DIP), one of the most essential obstacles is the image denoise algorithm by the reason of a very large demand of high quality noise-free images therefore there are many image denoise algorithms have been invented in the time of two decades. Bilateral filter is one of the most impressive and feasible algorithms, which is usually applied for denoise propose, but the performance of the Bilateral filter is substantially bank on three parameters: spatial variance, radiometric variance and window size. Consequently, this paper investigates the performance influence impact of spatial variance, radiometric variance, window size for the Bilateral Filter in the denoise propose. In the denoise experiment, Bilateral filter (BF) is applied on three noisy standard images under five Gaussian noise power levels and the best results in the PSNR prospective point of view from deniose algorithm is picked. Moreover, an optimal value of three parameters: spatial variance, radiometric variance, window size, which make the performance of Bilateral filter the highest PSNR, are extensively investigated for each types of tested images and each noise powers.
-
ItemComparative Experimental Exploration of Robust Norm Functions for Iterative Super Resolution Reconstructions under Noise Surrounding( 2015) Vorapoj PatanavijitIn DIP (Digital Image Processing) research society, the multi-frame SRR (Super Resolution Reconstruction) algorithm has grown to be the momentous theme in the last ten years because of its cost e ectiveness and its superior spectacle. Consequently, for a multi-frame SRR algorithm which is commonly comprised of a Bayesian ML (Maximum Likelihood) approach and a regularization technique into the unify SRR framework, numerous robust norm functions (which have both redescending and non-redescending in uence functions) have been commonly comprised in the unify SRR framework for increasingly against noise or outlier. First, this paper presents the mathematical model of several iterative SRR based on a Bayesian ML (Maximum Likelihood) approach and a regularization technique. Three groups of robust norm functions (a zero-redescending in uence function (Tukey's Biweight, Andrew's Sine and Hampel), a nonzero-redescending in uence function (Lorentzian, Leclerc, Geman&McClure, Myriad and Meridian) and a non-redescending in uence function (Huber)) are mathematically incorporated into the SRR framework. The close form solutions of the SRR framework based on these robust norm functions have been concluded. Later, the experimental section utilizes two standard images of Lena and Susie (40th) for pilot studies and fraudulent noise patterns of noiseless, AWGN, Poisson, Salt&Pepper, and Speckle of several magnitudes used to contaminate these two standard images. In order to acquire the maximum PSNR, the comparative experimental exploration has been done by comprehensively tailoring all experimental parameters such as step-size, regularization parameter, norm constant parameter.
-
ItemComputational Tutorial of Steepest Descent Method and Its Implementation in Digital Image Processing( 2013) Vorapoj PatanavijitIn the last decade, optimization techniques have extensively come up as one of principal signal processing techniques, which are used for solving many previous intractable problems in both digital signal processing (DSP) problems and digital image processing (DIP) problem. Due to its low computational complexity and uncomplicated implementation, the Gradient Descent (GD) method [1] is one of the most popular optimization methods for problems, which can be formulated as a differentiable multivariable functions. The GD method is ubiquitously used from basic to advanced researches. First, this paper presents the concept of GD method and its implementations for general mathematical problems. Next, the computation of GD processes is shown step by step with the aim to understand the effect of important parameters (such as its initial value and step size) to the performance of GD. Later, the computational concept of GD method for DIP problems [2-5] is formulated and the computation of GD is demonstrated step by step. The effect of the initial value and the step size to the performance of GD method in DIP is also presented.
-
ItemA Conceptual Framework of Super Resolution Reconstruction (SRR) Techniques( 2016) Vorapoj PatanavijitTypically, a spatial resolution (Pixel per Area) is an important factor used to define the image quality. Due to the dramatically advance of digital image processing in this decade, high resolution (HR) images are in demand because HR images give more detail and information that directly impact their application performance. Today, there are several techniques that can capture high resolution images such as resolution increment by reducing pixel side. Consequently this high resolution sensor is so expensive and do not proper for general applications. Moreover, due to reducing pixel side, the SNR of sensors decrease. From the signal processing theory, the alternative algorithm for increasing resolution of captured image is called “Super Resolution Reconstruction or SRR” that can solve this problem. Hence, the SRR refers that the process of increasing resolution and improving the quality of image to be higher resolution and better quality. This paper aim to review the ideal and concept of the SRR technique and its SRR observation model but this paper don’t review all SRR frameworks because there are so many proposed SRR techniques. Author hopes that the SRR ideal and concept framework reviewed in this paper will motivate the reader to conduct in this research areas.
-
ItemConceptual Framework of Super Resolution Reconstruction Based on Frequency Domain From Aliased Multi-Low Resolution Images: Theory Part( 2016) Vorapoj PatanavijitTypically, a spatial resolution (Pixel per Area) is an important factor used to define the image quality. Due to the dramatically advance of digital image processing in this decade, high resolution (HR) images are in demand because HR images give more detail and information that directly impact their application performance. Today, there are several techniques that can capture high resolution images such as resolution increment by reducing pixel side. Consequently this high resolution sensor is so expensive and do not proper for general applications. Moreover, due to reducing pixel side, the SNR of sensors decrease. From the signal processing theory, the alternative algorithm for increasing resolution of captured image is called “Super Resolution Reconstruction or SRR” that can solve this problem. Hence, the SRR refers that the process of increasing resolution and improving the quality of image to be higher resolution and better quality. This paper aim to review the ideal and concept of the SRR technique and its SRR observation model but this paper don’t review all SRR frameworks because there are so many proposed SRR techniques. Author hopes that the SRR ideal and concept framework reviewed in this paper will motivate the reader to conduct in this research areas.
-
ItemEmpirical Exploration Achievement of Noise Removal Algorithm Based on Trilateral Filter for Both Gaussian and Impulsive Noise Ambiance( 2016-06) Vorapoj PatanavijitAlthough the Bilateral filter is one of the most realistic and virtuoso noise removal algorithms, which is often proposed for Gaussian noise in 1998, the Bilateral filter (BF) ineffectively works under the impulsive noise. Consequently, Trilateral filter (which is a modification Bilateral filter) was first proposed by Roman Garnett et al. in 2005 and this filter is based on the hybrid consisting of Bilateral filter and Rank-Ordered Absolute Differences (ROAD) statistic for automatically attenuating or excluding of Gaussian and impulsive noise. Thereby, this research paper empirically explores the efficient influence impact of these four parameters (spatial, radiometric, ROAD and joint impulsivity variance) of the Trilateral filter (TF) when this Trilateral filter (TF) is used for noise removal prospective attitude. In the noise removal exploration, Trilateral filter (TF) is used for five noisy standard images (Girl-Tiffany, Pepper, Baboon, House and Resolution) under five Gaussian noises and five Impulse noise, compared with state-of-the-art algorithms such as Bilateral filter (BF) and median filter. Subsequently, the highest result in the PSNR prospective attitude is nominated. Supplementary, an empirically exploration optimal value of ROAD variance and joint impulsivity variance that yield the highest PSNR is empirically explored for each images and each noise cases.
-
ItemExperimental Analysis of Performance Comparison on Both Linear Filter and Bidirectional Con dential Technique for Spatial Domain Optical Flow Algorithm( 2013) Darun Kesrarat ; Vorapoj PatanavijitOptical flow is a method for classifying the density velocity or motion vector (MV) in a degree of pixel basis for motion classification of image in video sequences. In actual situation, many unpleasant situations usually generate noises over the video sequences. These unpleasant situations corrupt the performance in efficiency of optical flow. In turn to increase the efficiency of the MV, this research work proposes the performance comparison on linear filter and bidirectional confidential technique for spatial domain optical flow algorithms. Our experiment concentrates on the 3 classical spatial based optical flow algorithms (such as spatial correlation-based optical flow (SCOF), Horn-Schunk algorithm (HS) and Lucas- Kanade algorithm (LK)). Different standard video sequences such as AKIYO, CONTAINER, COASTGUARD, and FOREMAN are comprehensively evaluated to demonstrate the effectiveness results. These video sequences have differences in aspect of action and speed in foreground and background. These video sequences are also debased by the Additive White Gaussian Noise (AWGN) at different noise degree (such as AWGN at 25 dB, 20 dB, and 15 dB consequently). Peak Signal to Noise Ratio (PSNR) is utilized as the performance index in our observation.
-
ItemExperimental Efficiency Analysis in Robust models of Spatial Correlation Optical Flow Methods under Non Gaussian Noisy Contamination( 2013-05) Darun Kesrarat ; Vorapoj PatanavijitIn this paper, we present a performance analysis of several robust models of spatial correlation optical flow algorithms including an original spatial correlation optical flow (SCOF), bidirectional for high reliability optical flow (BHR), gradient orientation information for robust motion estimation (GOI), and robust and high reliability based on bidirectional symmetry and median motion estimation (RHR) under the non Gaussian noise conditions. The simulated results are tested on 4 different in foreground and background movement characteristics of standard sequences (AKIYO, CONTAINER, COASTGUARD, and FOREMAN) in a degree of 0.5 sub-pixel translation. In our experiment, an original sequence (no noise), and noise contaminated sequences on Salt & Pepper Noise (SPN) at density (d) = 0.025d, and 0.005d, Speckle Noise (SN) at variance (v) = 0.05v, and 0.01v, and Poisson Noise (PN) are utilized. The experiment concentrates on Peak Signal to Noise Ratio (PSNR) as an indicator in the experimental performance analysis.
-
ItemAn Experimental Performance Analysis of Image Reconstruction Techniques under Both Gaussian and Non-Gaussian Noise Models( 2012-07) Vorapoj PatanavijitRecently, 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.
-
ItemExperimental Study Efficiency of Robust Models of Lucas-Kanade Optical Flow Algorithms in the Present of Non-Gaussian Noise( 2012-07) Vorapoj PatanavijitThis paper presents experimental efficiency study of noise tolerance model of spatial optical flow based on LucasKanade (LK) algorithms such as original LK with kernel of Barron, Fleet, and Beauchemin (BFB), confidence based optical flow algorithm for high reliability (CRR), robust motion estimation methods using gradient orientation information (RGOI), and a novel robust and high reliability for LucasKanade optical flow algorithm using median filter and confidence based technique (NRLK) under several NonGaussian Noise. These experiment results are comprehensively tested on several standard sequences (such as AKIYO, COASTGUARD, CONTAINER, and FOREMAN) that have differences speed, foreground and background movement characteristics in a level of 0.5 sub-pixel displacements. Each standard sequence has 6 sets of sequence included an original (no noise), Poisson Noise (PN), Salt&Pepper Noise (SPN) at density (d) = 0.005 and d = 0.025, Speckle Noise (SN) at variance (v) = 0.01 and v = 0.05 respectively which Peak Signal to Noise Ratio (PSNR) is concentrated as the performance indicator.
-
ItemExperimental Study on Image Reconstruction from Spatial Correlation-based Optical Flow Motion Vector over Non Gaussian Noise Contamination using Reversed Confidential with Bilateral FilterIn motion estimation, noise is a verity to degrade the performance in optical flow for determining motion vector. This paper examines the performance of noise tolerance model in spatial correlation-based optical flow for image reconstruction from motion vector where the source sequences are contaminated by non Gaussian noise. There are Poisson Noise, Salt & Pepper Noise, and Speckle Noise. In the experiment, several standard sequences in different styles are used and the applied combination model of reversed confidential with bilateral filter on spatial correlation-based optical flow is mainly focused to determined the best condition to apply this model with. The result in image reconstruction from motion vector is used in performance comparison with traditional noise tolerance models by using Peak Signal to Noise Ratio (PSNR) as a primary index for studying.
-
ItemA fast image recovery using compressive sensing technique with block based orthogonal matching pursuit( 2009-12) Vorapoj Patanavijit ; Parichat Sermwuthisarn ; Supatana AuethavekiatTraditionally, the problems of applying Orthogonal Matching Pursuit (OMP) to large images are its high computing time and its requirement for a large matrix. In this paper, we propose a fast image recovery algorithm by dividing the image into block of nxn pixels and applying OMP to each nxn block instead of the entire image. The key idea is that small matrix requires less computing time and less memory. In the experiment, the block based OMP was applied to three standard test images: Lena, Mandrill and Pirate. Compared to standard OMP, block based OMP required less computing time while giving comparable PSNR.
-
ItemFast Image Restoration Technique for Car License Plate Based on PWS filter Using 2DPCA Algorithm( 2013) Vorapoj PatanavijitIn this paper, we propose the algorithm that used to identify car license plate that the capture images come in degraded version by combining a PWS filtering technique with a 2DPCA algorithm. From experiment results, our algorithm has three advantages. First, it can be operated to the image directly without transforming the structure of the image, which is two dimensional data, into a vector. Second, it can be implemented with less burden of computation and requires less memory. At last, less time is required to restore the image.
-
Item