Mean Median Filter Research Articles

Spatial domain noise removal filtering for low-resolution digital images

<p>n this research work, six different filters are applied on a low resolution 8 b/pixel gray-scale images, which operate on small sub-images (windows of 3×3 to 11×11 pixels). The enhanced images are used to compare the efficiency of the different six filters using the peak signal to noise ratio (PSNR) image quality measure. Noise peak elimination filter (PSNR)=36.63) outperforms others, such as median filter (PSNR=36.61), while corruption estimation (PSNR=36.03) significantly cuts processing time by only processing the corrupted pixels while maintaining image details. Mean filter (PSNR=34.05) is sensitive to outliers, which cause the image's sharpness and fine features to be lost. By avoiding averaging across edges, bimodal-averaging filter (PSNR=35.30), which improves on the mean filter, chooses the mean of the biggest population. The median-mean filtering (PSNR=36.32), which combines median and mean filters and determines the output pixel by averaging the median and some nearby pixels, is another improvement above averaging.</p>

Hybrid singular value decomposition based alpha trimmed mean-median filter in eliminating high density salt and pepper noise from grayscale image

Hybrid singular value decomposition based alpha trimmed mean-median filter in eliminating high density salt and pepper noise from grayscale image

Adaptive ground-roll attenuation using local nonlinear filtering

Source-generated coherent noise suppression is a long-standing issue in land seismic data processing. Traditional methods—such as f–k filtering, τ-p transform, and median filtering—often fail to remove coherent noise in irregular data, especially in the case that the noise highly overlaps with the signal. To address this issue, we developed a method based on local nonlinear filtering (LNF) to attenuate coherent noise in seismic data. In the proposed method, a novel similarity coefficient algorithm is presented to determine the optimal velocity of coherent noise. An initial model of coherent noise is constructed using a mixed median–mean filter and is then refined by the proposed weighting correction algorithm. After modeling the noise, coherent noise in seismic data can be then suppressed by adaptively subtracting the estimated noise from the raw seismic data. To verify the effectiveness of the proposed method, we applied it to regular and irregular synthetic data and a 3D field dataset containing strong ground roll. The results demonstrate that the new method successfully attenuates the ground roll and obtains better results in ground-roll attenuation and the preservation of useful signals compared with the three commonly used methods. Synthetic and real data examples indicate that the method is applicable to the attenuation of coherent noise in irregular data.

Image recognition based on Boltzmann machine algorithm

Due to the influence of recognition parameters, image recognition has low recognition accuracy, long recognition time and large storage cost. Therefore, an automatic image recognition method based on Boltzmann machine is proposed. Based on threshold method and fuzzy set method, image malformation correction is performed. The mean filter and median filter are combined to eliminate the influence of image filtering, and the pre-processing of image is completed by using the fuzzy enhancement of image. Based on the restricted Boltzmann method, the network model is dynamically evolved, and the identification parameters of each shape and contour are obtained. Different shapes and contours are classified and recognized. Simulation results show that image recognition method based on human-computer interaction has high recognition ability, shortens the time cost and greatly reduces the space needed for node storage.

Speckle noise reduction in coherent imaging systems via hybrid median–mean filter

Images recorded by coherent imaging systems, including laser-based photography, digital holography (DH), and digital holographic microscopy (DHM), are severely distorted by speckle noise. We present a single-shot image processing method to reduce the speckle noise, coined hybrid median–mean filter (HM2F), which is based on the average of conventional median-filtered images with different kernel sizes. The synergic combination of the median filter and mean approach provides a denoised image with reduced speckle contrast while the spatial resolution is kept at up to 97% of the original value. The HM2F method is compared with the conventional median filter approach, the 3D block matching filter, the nonlocal means filter, the 2D windowed Fourier transform filter, and the Wiener filter using different speckle-distorted images to benchmark its performance. Based on the experimental results and the simplicity of the technique, HM2F is proposed as an effective denoising tool for reducing the speckle noise in laser-based photography, DH, and DHM.

PERFORMANCE ANALYSIS OF NOISE REMOVAL TECHNIQUES FOR FACIAL IMAGES- A COMPARATIVE STUDY

Image de-noising has been a challenging issue in the field of digital image processing. It involves the manipulation of image data to produce a visually high quality image. While maintaining the desired information in the quality of an image, elimination of noise is an essential task. Various domain applications such as medical science, forensic science, text extraction, optical character recognition, face recognition, face detection etc. deal with noise removal techniques. There exist a variety of noises that may corrupt the images in different ways. Here, we explore filtering techniques viz. Mean filter, Median filter and Wiener filter to remove noises existing in facial images. The noises of our interest are namely; Gaussian noise, Salt & Pepper noise, Poisson noise and Speckle noise in our study. Further, we perform a comparative study based on the parameters such as Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Structure Similarity Index Method (SSIM). For this research work, MATLAB R2013a on Labeled faces in Wild (lfw) database containing 120 facial images is used. Based upon the aforementioned parameters, we have attempted to analyze the performance of noise removal techniques with different types of noises. It has been observed that MSE, PSNR and SSIM for Mean filter are 44.19 with Poisson noise, 35.88 with Poisson noise and 0.197 with Gaussian noise respectively whereas for that of Median filter, these are 44.12 with Poisson noise, 46.56 with Salt & Pepper noise and 0.132 with Gaussian noise respectively. Wiener filter when contaminated with Poisson, Salt & Pepper and Gaussian noise, these parametric values are 44.52, 44.33 and 0.245 respectively. Based on these observations, we claim that the Median filtering technique works the best when contaminated with Poisson noise while the error strategy is dominant. On the other hand, Median filter also works the best with Salt & Pepper noise when Peak Signal to Noise Ratio is important. It is interesting to note that Median filter performs effectively with Gaussian noise using SSIM.

Denoising Terahertz Image Using Non-Linear Filters

This work was supported by the National Natural Science Foundation of China, (Grant No. 11872058), the Sichuan Science and Technology Program of China (No.2019YFG0114) Abstract A major challenge in the processing and analysis of images is the presence of noise especially in terahertz images. Denoising techniques are designed to remove noise or distorted images while maintaining the original image quality. In this paper, terahertz image denoising is proposed using different filtering methods. This work provides an algorithm that denotes the terahertz image with a non- linear and spatial function of mean and medium filter. The THz images were transformed using Gaussian and Salt & Pepper noise at different percentage levels from 5% to 50%. Universal Image Quantitative Analysis was utilized specifically the most renowned PSNR, MSE, MAE and IEF were explored in this work. Experimental results show that the mean-median filter outperforms Salt and Pepper in removing the Gaussian noise especially when noise factor is increasing. Keywords: Mean- median filter, terahertz image, Gaussian noise, Salt & pepper noise DOI: 10.7176/CEIS/12-2-02 Publication date: April 30 th 2021

A comparative study of time of flight extraction methods in non-line-of-sight location

Non-line-of-sight location is an active detection technology which is used to detect the position of objects out of sight by extracting the time of flight. It is a research hotspot in recent years. In order to study the performance differences of mean filter, median filter and Gaussian filter in extracting time of flight, firstly, the energy changing model of photon flight model is optimized by photometry, and then the parameters of the three filtering methods are optimized and analyzed. After that, the adaptability of these three extraction methods to the maximum value judgment method and probability threshold weighted judgment method is analyzed. Finally, the accuracy and stability of these three time extraction algorithms are compared by using the positions of devices and invisible object as variables. The simulation results show that the median filter is suitable for a narrow environment and it has the high accuracy in positioning; the locations with Gaussian filter have good positioning stability and there is a wider selection range of filtering parameters when the signal is processed with Gaussian filter.

Novel Air Temperature Measurement Using Midwave Hyperspectral Fourier Transform Infrared Imaging in the Carbon Dioxide Absorption Band

Accurate visualization of air temperature distribution can be useful for various thermal analyses in fields such as human health and heat transfer of local area. This paper presents a novel approach to measuring air temperature from midwave hyperspectral Fourier transform infrared (FTIR) imaging in the carbon dioxide absorption band (between 4.25–4.35 μm). In this study, the proposed visual air temperature (VisualAT) measurement is based on the observation that the carbon dioxide band shows zero transmissivity at short distances. Based on analysis of the radiative transfer equation in this band, only the path radiance by air temperature survives. Brightness temperature of the received radiance can provide the raw air temperature and spectral average, followed by a spatial median-mean filter that can produce final air temperature images. Experiment results tested on a database obtained by a midwave extended FTIR system (Telops, Quebec City, QC, Canada) from February to July 2018 show a mean absolute error of 1.25 ∘K for temperature range of 2.6−26.4 ∘C.

Segmentation and Boundary Detection of Fetal Kidney Images in Second and Third Trimesters Using Kernel-Based Fuzzy Clustering

Organ segmentation is an important step in Ultrasound fetal images for early prediction of congenital abnormalities and to estimate delivery date. In many applications of 2D medical imaging, they face problems with speckle noise and object contours. Frequent scanning of fetal leads to clinical disturbances to the fetal growth and the quantitative interpretation of Ultrasonic images also a difficult task compared to other image modalities. In the present work a three-stage hybrid algorithm has been developed to segment the US fetal kidney images for the detection of shape and contour. At the first stage the hybrid Mean Median (Hybrid MM) filter is applied to reduce the speckle noise. Then a kernel based Fuzzy C - means clustering is used to detect the shape and contour. Finally, the texture features are obtained from the segmented images. Based on the obtained texture features, the abnormalities are detected. The Gaussian Radial basis function provides an accuracy of 80% at the second and third trimesters with weighted constant ranging from 4 to 8, compared to other global kernel functions. Similarly the proposed method has an accuracy of 86% with compared to other FCM techniques.

Thoracic hybrid PET/CT registration using improved hybrid feature intensity multimodal demon

PET/CT images are commonly used for cancer detection and management. In order to obtain high accuracy of tumour interpretation as well as the following treatment planning, accurate registration of these two is highly important. The use of hybrid PET/CT scanner to physically co-register these images does not fully solve the nonlinear misregistration problem due to physiological motions such as breathing and cardiac motions. This paper proposes an improved hybrid feature intensity multimodal demon registration algorithm to register PET/CT images acquired from hybrid PET/CT scanner. In basis, the method comprises of 3 major steps: PET sinogram filtering based on 3D hybrid mean-median filter; segmentation of structures from both PET/CT images (the lungs, the heart and the tumour) and registration of PET/CT images using feature intensity multimodal demon method. The segmented structures acted as “features” to be integrated in the registration process in which an improved intensity based multimodal demon registration was used. The improved multimodal demon used a combination of sum of conditional variations (SCV) and multimodality independent neighbourhood descriptive (MIND) similarity measures. Overall, the proposed hybrid feature intensity multimodal registration method was tested on the simulated NCAT PET/CT images and 21 clinical hybrid PET/CT datasets. Experimental results showed that the combination of SCV and MIND similarity measures produced the best registration result for PET/CT misregistration problem. The success of the registration of the patient datasets was validated through improved lung volume overlap between the PET and the CT lung post registration according to Dice coefficient as well as Hausdorff distance score. The registration method increased the Dice and the Hausdorff distance measures by 4.46% and 22.38% in average respectively after registration in which the registration improvement is more than double as achieved by FFD and previous method. It is concluded that our proposed PET/CT registration algorithm is better than the standard FFD registration technique; thus has the potential to be utilised in clinical application.

Stationary wavelet transform based radiometric error correction technique for NOAA-AVHRR sensor data

Due to the inaccuracy of the sensing devices, remote sensing images contain radiometric errors, which can be severe in many cases. Therefore, the preprocessing is an inevitable step in the remote sensing image analysis. This paper presents radiometric errors and evaluates methodologies to retrieve information contained in images by means of filtering in the spatial domain and wavelet domain. Among those, the wavelet techniques are more effective to reduce noise because of their ability to capture the energy of a signal in fewer wavelet coefficients. In this study, Stationary Wavelet Transform (SWT) method and its application to NOAA -18, 19 AVHRR/3 channel 3 and channel 4 images to correct radiometric error is presented. Qualitative and quantitative analysis was carried to evaluate the performance of SWT method, both by measuring the Peak Signal to Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM), mean value, standard deviation (SD) and by visual inspection. The SWT based method can remove radiometric errors effectively and preserves radiometric information to a desirable amount. From the results, SWT based method is better in smoothness and accuracy than the conventional mean filter, median filter and Discrete Wavelet Transform (DWT) based method.

Review on the Performance of Different Types of Filter in the Presence of Various Noises

An efficient filter must have the quality of removing noise at the same time preserving the information. On the other hand, noise contaminates in the field of signal processing and communication due to various unwanted nonlinear effects. So study of the characteristics of different types of noise is necessary. Also, how various types of filters response to these noises are also important? In this paper, we have studied the characteristics of four types of noise namely Salt & pepper noise, Gaussian noise, Speckle noise and Poisson noise and analyzed the performance of four basic types of filters such as Mean filter, Median filter, Wiener filter and Gaussian filter to suppress the noises mentioned previously.

Low-density noise removal based on lambda multi-diagonal matrix filter for binary image

Binary image denoising is a well-known problem, and it is the concern of diverse application areas. Many classical denoising techniques have evolved over the years, such as mean filter, median filter, and morphological filter. A new denoising method based on the multiplication of lambda multi-diagonal binary matrix (\(\lambda \)-MDBM) is proposed for binary images in this paper. In proposed method, first the users need to choose an appropriate lambda value from the set \(\{\lambda |\lambda \in [0.5,1]\}\), and then the hybrid noisy binary image matrix can be obtained by adding the noisy binary image matrix times the \(\lambda \)-MDBM and the transpose of the noisy binary image matrix times the \(\lambda \)-MDBM, and finally the de-noised binary image can be obtained by a preset threshold. The experimental results show that a new denoising method based on \(\lambda \)-MDBM is possible in peak signal-to-noise ratio and mean square error.

PET Image Reconstruction Incorporating 3D Mean-Median Sinogram Filtering

Positron Emission Tomography (PET) projection data or sinogram contained poor statistics and randomness that produced noisy PET images. In order to improve the PET image, we proposed an implementation of pre-reconstruction sinogram filtering based on 3D mean-median filter. The proposed filter is designed based on three aims; to minimise angular blurring artifacts, to smooth flat region and to preserve the edges in the reconstructed PET image. The performance of the pre-reconstruction sinogram filter prior to three established reconstruction methods namely filtered-backprojection (FBP), Maximum likelihood expectation maximization–Ordered Subset (OSEM) and OSEM with median root prior (OSEM-MRP) is investigated using simulated NCAT phantom PET sinogram as generated by the PET Analytical Simulator (ASIM). The improvement on the quality of the reconstructed images with and without sinogram filtering is assessed according to visual as well as quantitative evaluation based on global signal to noise ratio (SNR), local SNR, contrast to noise ratio (CNR) and edge preservation capability. Further analysis on the achieved improvement is also carried out specific to iterative OSEM and OSEM-MRP reconstruction methods with and without pre-reconstruction filtering in terms of contrast recovery curve (CRC) versus noise trade off, normalised mean square error versus iteration, local CNR versus iteration and lesion detectability. Overall, satisfactory results are obtained from both visual and quantitative evaluations.

A PMMW image denoising based on adaptive manifolds and high-dimensional mean median filter

Passive millimeter wave (PMMW) imaging which owes a lot of advantages over infa-red and optical imaging has been widely applied in many fields. But due to the confinement of the antenna aperture size, the acquired PMMW image is often noisy and in low resolution. A novel PMMW image denoising algorithm which combines the high-dimensional mean median filter and the adaptive manifolds is proposed in this paper. Our method firstly applies the weighted traversal algorithm to the input image for generating a high-dimensional image data. To accelerate the algorithm, the PCA based dimension reduction is performed subsequently to extract the main information and. Then the PCA retained high-dimensional data is processed by the proposed mean median filter which computes the mean of pixels near the centered median pixel in each dimension. The adaptive manifolds are generated based on the input image and the PCA retained image data. Finally, the adaptive manifolds filter is used to produce the final response. The experiment results indicate that our method has a good ability in removing the noise with edge preserving and is efficient for both the simulated images and the real PMMW image.

A switching median–mean filter for removal of high-density impulse noise from digital images

A new, efficient switching median–mean filter is proposed to remove high density impulse noise from digital images. The proposed method consists of detection and filtering stages. The pixels that are labeled noise-free in the detection stage remain unchanged and the noisy-pixels are replaced by the reference image based the proposed filter with 3×3 window. The simulation results show that the proposed filter outperforms some existing methods, both in vision and quantitative measurements. Moreover, our method is not only efficient to remove high density impulse noise and preserve the details in the image, but also possess a short processing time.

A Novel Mean Median Filter for Noise and Artifacts Suppression from Digital Images

A novel mean-median filter is proposed for the suppression of impulse noise and various artifacts from the digital images. Leading Diagonal Sorting Algorithm is used with the fixed 3x3 size working window to compute the median. Truncated mean is computed by defining the boundaries and truncating the pixel values in the filtering window that fall outside the defined boundary. Noise detection is carried in two steps: In the first step the ‘reference pixel’ is tested for the presence of impulses with Min-Max detection strategy. In the second stage of detection, edge preserving unique criteria is employed to further classify the pixel under test as ‘noisy’ or ‘edge belonged’. This intelligent edge preserving decision criterion decides whether the test pixel deserves the restoration or not and facilitates the restoration of the noisy pixel either by the ‘truncated mean’ or by the ‘window median’ based on the decision threshold. Performance of the proposed filter algorithm is studied on a large number of images with varying amounts of salt and pepper noise and several types of artifacts and their combinations. Simulation results prove that the proposed algorithm (PA) effectively suppresses the high density salt and pepper noise (SPN) and in addition, it performs excellent in suppressing image artifacts such as strip lines (both white and black), drop lines (both white and black), missing bands and noise blotches. The novelty of the proposed mean-median filter lies in the fact that it attempts to simultaneously suppress impulsive noise and artifacts with good edge preservation (as the maximum size of the filtering window is restricted to ‘3×3’ only) and also because a two stage detection mechanism is employed. The goodness of the proposed algorithm lies in replacing several independent filter schemes required for suppression of noise and artifacts of several types without much distorting the vital features of the image under test. The performance evaluation of the proposed algorithm is done in terms visual appearance and quantitative measures such as Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Image Enhancement Factor (IEF) and the Computation Time (CT). Simulation results are compared with other state of art algorithms to derive the meaningful conclusions. General Terms Image Processing, Noise and Artifacts Suppression.

An Improved Filter Method of Medical Image

The paper analysis mean filter, median filter and morphology filter. The paper studied order morphology theory and its filter character. Based on this, the paper proposed multi-scale and multi-structure filter algorithm based on order morphology. The result of simulation shows that the algorithm can not only keep the edge and detail of medical image but also can enhance the contrast of recovery image.

An efficient three-stage approach for removing salt & pepper noise from digital images

In this paper, an efficient three-stage scheme for the removal of salt & pepper noise based on an efficient impulse detector, an adaptive mean-median filter and the total variation inpainting method was proposed. This approach removes salt & pepper noise by detecting, estimating and modifying noisy pixels in an image. If a pixel is classified as noise, its intensity is first estimated by a mean-median filter; and then an accurate estimation is obtained through the use of the total variation inpainting method. Conversely, the pixel value is kept unchanged when it is classified as noise-free, yielding the quality of the restored image being well maintained. Experimental results demonstrate that the proposed three-stage approach can not only remove salt & pepper impulse noise efficiently, but also preserve the detailed information of an image well.