Hybrid Median Filter Research Articles

Efficient VLSI implementation of modular neural network based hybrid median filter

Efficient VLSI implementation of modular neural network based hybrid median filter

Improved detection of dry age-related macular degeneration from optical coherence tomography images using adaptive window based feature extraction and weighted ensemble based classification approach

BackgroundDry Age-related macular degeneration (AMD), which affects the older population, can lead to blindness when left untreated. Preventing vision loss in elderly needs early identification. Dry-AMD diagnosis is still time-consuming and very subjective, depending on the ophthalmologist. Setting up a thorough eye-screening system to find Dry-AMD is a very difficult task. MethodologyThis study aims to develop a weighted majority voting (WMV) ensemble-based prediction model to diagnose Dry-AMD. The WMV approach combines the predictions from base-classifiers and chooses the class with greatest vote based on assigned weights to each classifier. A novel feature extraction method is used along the retinal pigment epithelium (RPE) layer, with the number of windows calculated for each picture playing an important part in identifying Dry-AMD/normal images using the WMV methodology. Pre-processing using hybrid-median filter followed by scale-invariant feature transform based segmentation of RPE layer and curvature flattening of retina is employed to measure exact thickness of RPE layer. ResultThe proposed model is trained on 70% of the OCT image database (OCTID) and evaluated on remaining OCTID and SD-OCT Noor dataset. Model has achieved accuracy of 96.15% and 96.94%, respectively. The suggested algorithm's effectiveness in Dry-AMD identification is demonstrated by comparison with alternative approaches. Even though the suggested model is only trained on the OCTID, it has performed well when tested on additional dataset. ConclusionThe suggested architecture can be used for quick eye-screening for early identification of Dry-AMD. The recommended method may be applied in real-time since it requires fewer complexity and learning-variables.

An Efficient USE-Net Deep Learning Model for Cancer Detection

Breast cancer (BrCa) is the most common disease in women worldwide. Classifying the BrCa image is extremely important for finding BrCa at an earlier stage and monitoring BrCa during treatment. The computer-aided detection methods have been used to interpret BrCa and improve the detection of BrCa during the screening and treatment stages. However, if a new BrCa image is generated for the treatment, it will not classify correctly. The main objective of this research is to classify the BrCa images for newly generated images. The model performs preprocessing, segmentation, feature extraction, and classification. In preprocessing, a hybrid median filtering (HMF) is used to eliminate the noise in the images. The contrast of the images is enhanced using quadrant dynamic histogram equalization (QDHE). Then, ROI segmentation is performed using the USE-Net deep learning model. The CaffeNet model is used for feature extraction on the segmented images, and finally, classification is made using the improved random forest (IRF) with extreme gradient boosting (XGB). The model obtained 97.87% accuracy, 98.45% sensitivity, 95.24% specificity, 98.96% precision, and 98.70% f1-score for ultrasound images. The model gives 98.31% accuracy, 99.29% sensitivity, 90.20% specificity, 98.82% precision, and 99.05% f1-score for mammogram images.

Efficient adaptive enhanced adaboost based detection of spinal abnormalities by Machine learning approaches

Spinal abnormalities are commonly occurring disorders that are caused by injuries, osteoporosis (benign) and neoplastic infiltration (malignant). Patients with one of the listed malignancies must undergo therapy as soon as possible to prevent the progress of the disease and to avoid further bone damage to preserve a better quality of life. Classification of patients into low or high-risk groups is an important step in diagnosing a disorder, which led biomedical and bioinformatics research teams to investigate the usage of machine learning (ML) technologies. In this paper, ML techniques are adapted to a public lumbar spine dataset to detect vertebral fractures. First, the dataset was preprocessed by using the contour-based hybrid median filter with histogram equalization. Then the Mask LSTM-based R.O.I. segmentation techniques are applied to segment the spinal images. Finally, the suggested stacked Adaptive Enhanced AdaBoost (AE-AdB) gets trained on whole images to enhance the accuracy of the data classification. Results of the experiment show that the attained accuracy (97.86%) of the AE-AdB classifier was significantly higher than that of the other classifiers, namely Convolution Neural Network (CNN) (74%) and Extreme Gradient Boosting Algorithm (XGBoost) (84%).

Monitoring and Analysis of Agricultural Field Parameters in Order to Increase Crop Yield through a Colored Object Tracking Robot, Image Processing, and IOT

Adequately watering plants is a challenging task. Over- and under-watering may harm plants and seeds, as excess or restraint watering reduces crop production and yield. This study presents a method to remotely monitor and efficiently water agricultural fields to increase crop production by utilizing advanced technologies such as internet things, robotics, image processing, and neural networks. Accurate smoothing and image segmentation techniques were employed to study the plants' conditions. Color median, Gaussian, and hybrid median filters were employed to preprocess the data before segmentation and classification. The hybrid median filter and multilevel luminance grading system were employed to increase the quality of the image. The k-means clustering approach was used for image segmentation. The signal-to-noise ratios of the original and recreated images were compared and analyzed.

Compression of Medical Images Using Wavelet Transform and Metaheuristic Algorithm for Telemedicine Applications

Medical image compression becomes necessary to efficiently handle huge number of medical images for storage and transmission purposes. Wavelet transform is one of the popular techniques widely used for medical image compression. However, these methods have some limitations like discontinuity which occurs when reducing image size employing thresholding method. To overcome this, optimization method is considered with the available compression methods. In this paper, a method is proposed for efficient compression of medical images based on integer wavelet transform and modified grasshopper optimization algorithm. Medical images are pre-processed using hybrid median filter to discard noise and then decomposed using integer wavelet transform. The proposed method employed modified grasshopper optimization algorithm to select the optimal coefficients for efficient compression and decompression. Four different imaging techniques, particularly magnetic resonance imaging, computed tomography, ultrasound, and X-ray, were used in a series of tests. The suggested method's compressing performance is proven by comparing it to well-known approaches in terms of mean square error, peak signal to noise ratio, and mean structural similarity index at various compression ratios. The findings showed that the proposed approach provided effective compression with high decompression image quality.

Detection and Diagnosis of Brain Tumor Using Wavelet Transform and Machine Learning Model

Brain tumor diagnosis using magnetic resonance images (MRI) is one of the hot topics of interest in medical field. In this paper, a novel approach for brain tumor diagnosis is proposed. Four major phases are involved in the proposed approach which are preprocessing, tumor segmentation, feature extraction and classification. Hybrid median filtering method is implemented to suppress noise. Otsu thresholding method is used for isolating tumor from its background tissues. Wavelet transform is adopted for feature extraction. Support vector machined is used for classifying the brain tumors. Classification potential of the proposed system is tested on Harvard and BRATS data set.

Performance analysis of improved hybrid median filter applied to X-ray computed tomography images obtained with high-resolution photon-counting CZT detector: A pilot study

We evaluated the performance of an improved hybrid median filter (IHMF) applied to X-ray computed tomography (CT) images obtained using a high-resolution photon-counting cadmium zinc telluride (CZT) detector. To study how the proposed approach improves the image quality, we measured the noise levels and the overall CT-image quality. We established a CZT imaging system with a detector length of 5.12 cm and thickness of 0.3 cm and acquired phantom images. To evaluate the efficacy of the proposed filter, we first modeled two conventional median filters. Subsequently, we were able to achieve a normalized noise power spectrum result of ∼10−8 mm2, and furthermore, the proposed method improved the contrast-to-noise ratio by a factor of ∼1.51 and the coefficient of variation by 1.55 relative to the counterpart values of the no-filter image. In addition, the IHMF exhibited the best performance among the three filters considered as regards the peak signal-to-noise ratio and no-reference-based image-quality evaluation parameters. Thus, our results demonstrate that the IHMF approach provides a superior image performance over conventional median filtering methods when applied to actual CZT X-ray CT images.

MRI Brain Tumor Segmentation with Intuitionist Possibilistic Fuzzy Clustering and Morphological Operations

Digital Image Processing (DIP) is a well-developed field in the biological sciences which involves classification and detection of tumour. In medical science, automatic brain tumor diagnosis is an important phase. Brain tumor detection is performed by Computer-Aided Diagnosis (CAD) systems. The human image creation is greatly achieved by an approach namely medical imaging which is exploited for medical and research purposes. Recently Automatic brain tumor detection from MRI images has become the emerging research area of medical research. Brain tumor diagnosis mainly performed for obtaining exact location, orientation and area of abnormal tissues. Cancer and edema regions inference from brain magnetic resonance imaging (MRI) information is considered to be great challenge due to brain tumors complex structure, blurred borders, besides exterior features like noise. The noise compassion is mainly reduced along with segmentation stability by suggesting efficient hybrid clustering method merged with morphological process for brain cancer segmentation. Combined form of Median Modified Wiener filter (CMMWF) is chiefly deployed for denoising, and morphological operations which in turn eliminate nonbrain tissue, efficiently dropping technique’s sensitivity to noise. The proposed system contains the main phases such as preprocessing, brain tumor extraction and post processing. Image segmentation is greatly achieved by presenting Intuitionist Possibilistic Fuzzy Clustering (IPFC) algorithm. The algorithm’s stability is greatly enhanced by this clustering along with clustering parameters sensitivity reduction. Then, the post processing of images are done through morphological operations along with Hybrid Median filtering (HMF) for attaining exact tumors representations. Additionally, suggested algorithm is substantiated by comparing with other existing segmentation algorithms. The outcomes reveal that suggested algorithm achieves improved outcomes pertaining to accuracy, sensitivity, specificity, and recall.

Modified algorithms for signal nonlinear trend detection

There is a problem of nonlinear (abrupt) signal trend detection in many digital signals processing practical applications. In particular, in the field of biomedical signals processing, the actual task is the elimination of abrupt signal baseline distortions caused by the patient's movements. For processing such signals containing edges and other discontinues, linear filtering based on discrete Fourier or cosine transforms leads to significant smoothing of a signal. Median type algorithms related to nonlinear stable (robust) filters are successfully applied for filtering such signals, in particular, high efficiency is provided by median hybrid filters with finite impulse response (FIR). The article considers simple algorithms of the class of FIR-median hybrid filters used for signal nonlinear trend detection. It is proposed to modify these algorithms by replacing the operation of finding the median of the data in the sliding filter window with the calculation of their myriad, as well as adding weights (number of duplications) to certain window elements. Statistical estimates of filter efficiency according to the mean square error (MSE) criterion for test signals like “step” and “ramp” edges, and triangular peak and parabola have been obtained. The high efficiency of the investigated nonlinear filters for the listed test signals types and the improvements achieved as a result of the proposed filter modifications are shown based on the analysis of the filter output signals and statistical estimates of their quality. Some examples of processing biomedical signals of electroencephalograms which illustrate good quality of noise suppression and signal abrupt changes preservation, and motion artifacts removal without large signal distortions are given.

Effect of despeckling filters on the segmentation of ultrasound common carotid artery images

BackgroundCarotid intima-media thickness (IMT) measured in B-mode ultrasound image is an important indicator of Atherosclerosis disease. Speckle noise inherently present in ultrasounds’ thereby degrades the visual evaluation and limits the automated segmentation performance. The objective of this study is to investigate the effects of three despeckle filters on the segmentation of carotid IMT in ultrasound image. MethodsAutomated segmentation of IMT is achieved by utilizing fast fuzzy c-mean clustering and distance-regularized level set without re-initialization techniques. Manual segmentation has been done by an experienced radiologist. The performances of median, hybrid median and improved adaptive complex diffusion (IACDF) filters are examined and a quantitative and qualitative comparison among these filters has been reported on 151 DICOM images. Bland–Altman plots were used to compare IMT results of these filters.Furthermore, performances of above three filters are evaluated under different noise levels by individually adding speckle and salt and pepper noise in ten randomly selected images from 151 DICOM dataset. Plots between noise and quality evaluation metric parameters are used to compare de-noising performance of these filters. ResultsThe average processing time per image of proposed IMT measurement technique without-filter and with filter is approx 15.39 s max. ConclusionIt is shown that the median filter (window 5 × 5) measures better than hybrid median and IACDF filters. Finally, concluded that de-noising of ultrasound image before segmentation procedure certainly improves segmentation accuracy. Furthermore, it is observed that these filters do not impose serious computational burden and entail moderate processing time.

Denoising Performance Evaluation of Automated Age-Related Macular Degeneration Detection on Optical Coherence Tomography Images

Automated detection of eye diseases using artificial intelligence techniques on optical coherence tomography (OCT) images is widely researched in the field of ophthalmology. Such detections are usually performed with the aid of computers. Using high-level simulations, this study investigates and evaluates three automated age-related macular degeneration (AMD) detection flows in terms of computation time and detection accuracy for future hardware-accelerated designs of intelligent and portable OCT systems. In this study, a block-matching and 3-Dimension filter (BM3DF), a hybrid median filter (HMF), and an adaptive wiener filter (AWF) are used to denoise the OCT images. Support vector machine (SVM), AlexNet, GoogLeNet, and Inception-ResNet are employed for AMD detection. Moreover, Local binary patterns, linear configuration patterns, and transfer learning techniques are used to extract image features. Simulation results reveal that machine-learning-based automated AMD detection realizes a high detection accuracy of 95.91% accompanied by low computation time when using the HMF rather than the BM3DF. When considering deep-learning-based automated AMD detection, the combination of HMF and Inception-ResNet achieves the highest detection accuracy of 98.64% but is accompanied by a dramatic increase in computation time. However, only AlexNet achieves a detection accuracy of 96.40%, accompanied by low computation time. In this study, the tradeoffs between the computation time and detection accuracy have been revealed by comparing the denoising methods for the distinct automated AMD detections.

Study of hybrid median filter for the removal of various noises in digital image

The aim of this study is to construct a hybrid median filter using MATLAB and use that filter to remove impulse noise from an image, in order to produce an enhanced image. This is done in order to show how a hybrid median filter can be a more optimal option than a conventional median filter for removing impulse noise.

Blended noise suppression using a hybrid median filter, normal moveout and complex curvelet transform approach

The high-density acquisition technique can improve subsurface imaging accuracy. However, it increases production cost rapidly and limits the wide application in practice. To solve this issue, the high productivity blending acquisition technology has emerged as a promising way to significantly increase the efficiency of seismic acquisition and reduce production cost. The great challenge of the blending acquisition technology lies in the severe interference noise of simultaneous sources. Therefore, the success of the blending acquisition technology relies heavily on the effectiveness of separating effective energy from the blended noise. We propose a blended noise suppression approach by using a hybrid median filter, normal moveout (NMO), and complex curvelet transform (CCT) approach. First, median filter is applied to original data after NMO correction. Second, the CCT-based thresholding denoising method is used to extract the remained effective energy from the data after median filtering to get the preliminary de-blended result. Next, the updated data are obtained by subtracting the pseudo-de-blended data of the de-blended result from the original data, and the process iterates. Last, the final de-blended result is obtained by adding the retrieved energy at each iteration until the signal-to-noise ratio satisfies the desired level. We demonstrate the effectiveness of the proposed approach on simulated synthetic and field data examples.

A Segmentation Perspective of Non-Cooperative Iris Recognition

Iris Segmentation, an initial and vital stage of the iris recognition stage which directly affects the recognition accuracy. Especially, the non-cooperative environment that leads to contain many of the noise parameters in the captured image. Since the recognition accuracy of the iris biometrics system is extremely dependent on the proper iris segmentation, this paper is devoted to the segmentation perspective of the non-cooperative iris recognition system. The initial stage of the proposed method is started with applying a hybrid median filter algorithm to remove the possible noises and then a region-based level set algorithm is applied to overcome the identification of the concave property in the non-cooperative iris segmentation and enhanced Otsu’s thresholding method is applied to the pupil segmentation. UBIRIS, a publicly available iris database for the non-cooperative situation, Version 1 and Version 2 is used for the implementation purpose. The accuracy of the segmentation result is achieved as 94.56 and 94.53 for the UBIRISv.1 and UBIRISv.2 respectively which show the proposed method as a better one.

An effective non-cooperative iris recognition system using hierarchical collaborative representation-based classification

In recent years, non-cooperative iris recognition has gained a major role in biometric authentication system. However, owing to images captured in non-cooperative environments, it is quite a difficult job, for these images have specular reflections, blur, occlusions, and are off-axis. This research introduces an efficient non-cooperative iris recognition system developed for hierarchical collaborative representation-based classifier (HCRC). The proposed method includes three stages. The first stage involves image preprocessing. Initially, hybrid median filtering is done to reduce noise and to improve the image quality. Then, segmentation of the abnormal non-cooperative iris image is carried out by applying Geodesic Region-based Active Contour Level-set algorithm and threshold-based segmentation algorithm. In the second stage, 2 × 2 block-based Local Ternary Pattern (LTP) is applied to the segmented image. This gives upper and lower LTP histogram blocks. The final feature vector is obtained by combining these two blocks. In the third stage, the feature vectors are applied to the HCRC for classification on the basis of the iris database. The proposed iris recognition technique proved itself by achieving 98.60, 0.095 and 0.096 accuracy, false acceptance rate and false rejection rate, respectively.

Improved Color Attenuation Prior for Single-Image Haze Removal

This paper proposes a single image haze removal algorithm that shows a marked improvement on the color attenuation prior-based method. Through a vast number of experiments on a wide variety of images, it is discovered that there are problems in the color attenuation prior, such as color distortion and background noise, which arise due to the fact that the priors do not hold true in all circumstances. Successful resolution of these problems using the proposed algorithm shows its superior performance to other state-of-the-art methods in terms of both subjective visual quality and quantitative metrics, on both synthetic and natural hazy image datasets. The proposed algorithm also is computationally friendly, due to the use of an efficient quad-decomposition algorithm for atmospheric light estimation and a simple modified hybrid median filter for depth map refinement.

A 4K-Capable FPGA Implementation of Single Image Haze Removal Using Hazy Particle Maps

This paper presents a fast and compact hardware implementation using an efficient haze removal algorithm. The algorithm employs a modified hybrid median filter to estimate the hazy particle map, which is subsequently subtracted from the hazy image to recover the haze-free image. Adaptive tone remapping is also used to improve the narrow dynamic range due to haze removal. The computation error of the proposed hardware architecture is minimized compared with the floating-point algorithm. To ensure real-time hardware operation, the proposed architecture utilizes the modified hybrid median filter using the well-known Batcher’s parallel sort. Hardware verification confirmed that high-resolution video standards were processed in real time for haze removal.

Underwater visibility restoration using dehazing, contrast enhancement and filtering

A framework for underwater image haze removal and visibility enhancement is proposed. Hybrid median filter in conjunction with shock filter is utilized for initial enhancement of image. For further enhancement and color restoration, content adaptive detail enhancement is applied followed by contrast enhancement in the RGB and HSV color space. The proposed framework effectively dehazes the images, and also minimizes color distortion and artifacts. The quantitative and visual analysis effectiveness demonstrate the effectiveness of our proposed technique.

Wavelet Decomposition–Based Speckle Reduction Method for Ultrasound Images by Using Speckle-Reducing Anisotropic Diffusion and Hybrid Median

Ultrasound imaging is considered as a common facility of imaging techniques used in medical field because of its benefits, but the quality of ultrasound image is affected and degraded by the type of noise generated by the backscattered signals that interfere with each other and introduce a random interference pattern. This pattern is called speckle. In this article, we develop a new denoising method by introducing a speckle-reducing anisotropic diffusion filter into 4 sub-band images gained by using 2-dimensional wavelet decomposition. Moreover, according to frequency content on each sup band, we use hybrid median filter to get a full advantage of using nonlinear filtering in speckle noise. Finely to improve the quality of image and the overall performance, total variation is used in view of it is capability of brightness equality and edges preserving. The experimental result shows that the proposed method is highly effective in noise separation as well as edges preserving based on compression between linear scaling of the gray-level values despeckle filter, speckle-reducing anisotropic diffusion, median, hybrid median filter, and wavelet filters using image quality evaluation metrics.