Morphological Filtering Research Articles

Novel and fast EMD-based image fusion via morphological filter

Novel and fast EMD-based image fusion via morphological filter

Development of a Sharp-Tipped L-Shaped Stylus for Measurement of Nanoscale Sidewall Features

This paper presents a system for measuring nanoscale sidewall features using an L-shaped stylus with a sharp tip. By sharpening the tip of the stylus to the nanometer scale, it is possible to measure nanoscale sidewall features. The stylus shaft is deflected when the sharpened stylus tip contacts the measured surface, and this deflection is measured optically. In this study, we develop the fabrication method of an L-shaped sharpened stylus 20 μm in diameter and 4 mm in length by etching, CO2 laser processing, and bonding. Then, the measurement accuracy and repeatability are examined by measuring V-grooves with nanoscale features. The results clarify that the pitch is approximately 0.3 μm, which is almost the same as the value measured using the photographic image, but the depths measured by the styli fabricated by CO2 laser processing and bonding are approximately 0.15 and 0.07 μm, respectively. The depth measured by the stylus fabricated by bonding is smaller than that measured from the image (approximately 0.16 μm), presumably because of the morphological filter. The maximum repeatability errors for 10 measurements using the styli fabricated by CO2 laser processing and bonding are within ±0.056 and ±0.022 μm in scanning mode, respectively. It is also confirmed that the stylus tip exhibits minimal wear even after 500 measurements.

Selective weighted multi-scale morphological filter for fault feature extraction of rolling bearings

Morphological filtering shows effectiveness in vibration signal analysis because of its simplicity and efficiency. Considering that different structural elements have different effects on filtering results, a new multi-scale morphological filtering (MMF) method called selective weighted multi-scale morphological filter (SWMMF) is developed for integrating results of different scales based on adaptive weighting strategy. Firstly, four morphological operators (dilation–closing, closing–dilation, erosion–opening and opening–erosion) are integrated into a new combination difference morphological filter to strengthen effect of faulty component extraction. Secondly, this new morphological filter is further extended to multiple scales in order to overcome limitation of single scale filter. Finally, the filtered results of different scales are adaptively combined by using the whale optimization algorithm (WOA)-based selective weighting method. The effectiveness of multi-scale filter and selective weights is proved by comparing with single-scale and average weighting filter on simulation and real-world cases (bearing vibration signals with different defects). The testing results on vibration signals indicate that SWMMF is able to extract effectively defect frequency and the corresponding multiplication frequencies from bearing vibration signals with heavy noise. The testing results illustrate that SWMMF outperforms other representative MMFs (e.g., weighted multi-scale morphological gradient operator (WMMG), weighted multi-scale difference operator (WMDIF), weighted multi-scale average operator (WMAVG)) on impulsive feature extraction of bearing vibrations signals with various defects. Moreover, it is demonstrated that SWMMF has good applicability in bearing fault diagnosis due to setup of adaptive weights and selection of structure element.

Computed tomography (CT) Image-based analysis of concrete microstructure using virtual element method

The analysis of actual concrete microstructure is challenging because the microstructure has various length scales associated with small voids and various sizes of aggregate particles. To efficiently and effectively analyze complex microstructure like concrete, the image-based virtual element method (VEM) is presented. Concrete microstructure is reconstructed using the complementarity of X-ray and neutron computed tomography (CT) images with the morphological filtering. Based on the reconstructed microstructure, a VEM mesh is generated through integrating microstructural and homogeneous meshes. A microstructural mesh is directly obtained from the reconstructed microstructure, while a homogeneous mesh is employed to control the level of accuracy in the computational simulation. Then, the mesh refinement in the image-based VEM is easily performed through increasing the number of elements in a homogeneous mesh. Computational results converge to the reference solution of strain error under mesh refinement while capturing local stress and strain concentration.

Unsupervised change detection in SAR images based on generalized likelihood ratio test and a two-stage morphological filter

ABSTRACT Generalized likelihood ratio test (GLRT) is an efficient method to generate difference image (DI) for change detection (CD) using synthetic aperture radar (SAR) images. GLRT is usually applied with a fixed-size moving window to the neighbourhood regions in multitemporal SAR images. The fixed window may be however not optimal for all the pixels under test. To solve disadvantages of the GLRT method, an adaptive circular window is proposed in this work. While the adaptive square window and the best fixed square window achieve an average Kappa coefficient of 83.55% and 82.12%, respectively, the adaptive circular window improves the average Kappa coefficient by 83.65% in six datasets. As another difficulty, speckle noise reduces quality of DI in SAR change detection. Three steps are considered to minimize effects of the speckle noise: (1) a two-stage morphological filter is suggested to reduce the speckle noise; (2) to generate DI, the adaptive circular window for generalized likelihood ratio test (ACWGLRT) is proposed that reduces influence of the speckle noise while preserves the edge details of multitemporal images; and (3) spatial fuzzy c-means (SFCM) is used to reduce effects of the residual speckle-noise during DI classification. The experimental results show superior performance of the proposed change detection method with respect to several competitors. The proposed method has the best Kappa coefficient and percentage correct classification (PCC) in four datasets of Ottawa, San Francisco, Farmland C, and Inland water.

Abstract 2460: Assessing the effectiveness of transarterial chemoembolization using super-resolution ultrasound imaging and a rat model of hepatocellular carcinoma

Abstract Introduction: Hepatocellular carcinoma (HCC) remains a leading cause of cancer-related deaths worldwide. Transarterial chemoembolization (TACE) treatment of HCC restricts the blood supply to the tumor. Conventional assessment of tumor response is performed by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) or computed tomography (DCE-CT) at 4 to 6 weeks after the TACE procedure. An earlier indication of treatment effectiveness would improve patient management and prognosis. Super-resolution ultrasound (SR-US) imaging allows a ten-fold improvement in spatial resolution compared to traditional ultrasound (US) and allows visualization of microvascular networks. The goal of this preclinical research study was to use SR-US imaging to assess the effectiveness of TACE treatment in a relatively brief time frame. Methods: Eight male Sprague Dawley rats weighing around 300 g were injected with 5 million N1S1 cells in the upper left liver lobe and the tumors were allowed to grow for 14 days. The TACE procedure consisted of transhepatic arterial delivery of a mixture of 25 µL Lipiodol and 25 µL Doxorubicin (10 mg/kg) via a polyethylene microcatheter. After intravenous injection of a microbubble (MB) contrast agent, contrast-enhanced US images were acquired (N = 3000) with a preclinical system (Vevo 3100, FUJIFILM VisualSonics Inc) equipped with an MX201 linear array transducer. US images were processed to remove high motion frames, followed by tissue suppression filtering, and then MB localization and accumulation. Morphological filtering was used to enhance the vessel structures. In vivo US, MRI, and CT imaging were performed at baseline before TACE and again at 1 and 2 wk. After euthanasia, tumor tissue was removed for ex vivo analysis. Results: A rat model of HCC for assessing TACE treatment was introduced. Based on tumor size changes and residual perfusion after the TACE procedure, treatments were determined to be a complete responder, partial responder, or non-responder. Initial results demonstrate that SR-US imaging can sensitively detect any reduction in HCC perfusion as measured by a microvascular density (MVD) metric within 2 wk after a single TACE procedure. MVD measurements from the SR-US images were consistence with tumor volume data from MRI and CT imaging. Conclusions: In vivo SR-US images of tumor microvascular networks provided insight into treatment efficacy. Citation Format: Junjie Li, Katherine Brown, Megan Yociss, John Eisenbrey, Kenneth Hoyt. Assessing the effectiveness of transarterial chemoembolization using super-resolution ultrasound imaging and a rat model of hepatocellular carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2460.

Visual Scratch Defect Detection System of Aluminum Flat Tube Based on Cubic Bezier Curve Fitting Using Linear Scan Camera

This paper presents a scratch detection system based on a cubic Bezier curve fitting using a linear scan camera. The objective was to detect the scratch defects of an aluminum flat tube stably in real-time under complex uncertain background noise. To that end, according to the features of the input image of the linear scan camera and the scratch defects, the proposed method first segmented the input image to ten equal sections in a longitudinal direction, and for every section, OTSU thresholding and morphological filtering were used to reduce the background noise. After the image preprocessing, every section image was projected along a vertical direction to form a vertical histogram. After that, for each point of every vertical histogram, a novel curve fitting method based on the Monte Carlo method was employed to calculate the best fitted Bezier curve. All the curvatures of the middle point of the best fitted Bezier curves then formed a curvature curve, and the scratches were located by finding the peaks of the curvature curve. Next, the result of the ten sections were fused to find the final positions of the scratches. The experimental results based on the linear scan camera that captured the image of flat tubes on a moving speed of 2m/s showed that the proposed method can detect the scratch defects under complex background noise with a high success rate in real-time.

Acoustic-based whistle detection of drain hole for wind turbine blade

The blade is a crucial part of large-scale wind turbine for converting wind energy into electricity. Increasing attentions have been paid to the blade health monitoring in recent years. This paper focuses on acoustic-based surface damage detection of the blade and presents a novel intelligent detection method of the whistle produced by the drain hold using the pattern recognition. In the algorithm, a developed preprocessing strategy with multiband adaptive spectral subtraction can well reduce the random wind noise from raw acoustic signal. Moreover, a correlated empirical mode decomposition method in combination with morphological filtering is proposed for extracting time–frequency ridge features of whistle event. Finally, an incremental support vector machine based on adaptive reserved set strategy is designed for recognizing the whistle event. Experimental results demonstrate that proposed method is feasible and effective in whistle detection of drain hole.

Individual Maize Location and Height Estimation in Field from UAV-Borne LiDAR and RGB Images

Crop height is an essential parameter used to monitor overall crop growth, forecast crop yield, and estimate crop biomass in precision agriculture. However, individual maize segmentation is the prerequisite for precision field monitoring, which is a challenging task because the maize stalks are usually occluded by leaves between adjacent plants, especially when they grow up. In this study, we proposed a novel method that combined seedling detection and clustering algorithms to segment individual maize plants from UAV-borne LiDAR and RGB images. As seedlings emerged, the images collected by an RGB camera mounted on a UAV platform were processed and used to generate a digital orthophoto map. Based on this orthophoto, the location of each maize seedling was identified by extra-green detection and morphological filtering. A seed point set was then generated and used as input for the clustering algorithm. The fuzzy C-means clustering algorithm was used to segment individual maize plants. We computed the difference between the maximum elevation value of the LiDAR point cloud and the average elevation value of the bare digital terrain model (DTM) at each corresponding area for individual plant height estimation. The results revealed that our height estimation approach test on two cultivars produced the accuracy with R2 greater than 0.95, with the mean square error (RMSE) of 4.55 cm, 3.04 cm, and 3.29 cm, as well as the mean absolute percentage error (MAPE) of 3.75%, 0.91%, and 0.98% at three different growth stages, respectively. Our approach, utilizing UAV-borne LiDAR and RGB cameras, demonstrated promising performance for estimating maize height and its field position.

Estimation and evaluation of paddy rice canopy characteristics based on images from UAV and ground camera

<p id="C3">Phenotypic monitoring of rice in the field can be used to analyze traits related to rice yield, which is of great significance to guide rice cultivation management and yield prediction. In this study, to explore the applicability of image analysis methods to evaluate rice growth in different fields under multiple cultivars and cultivation environments, we estimated and evaluated the main phenotypic parameters of rice canopy in paddy fields with six different cultivars under three nitrogen treatments. Based on UAV and field fixed camera images, this study used image processing, three-dimensional modeling and machine learning to automatically estimate rice canopy coverage, plant height, and panicle number in the field, and evaluated the accuracy combining with the actual measurement results. The results showed that the rice canopy segmentation based on the decision tree classification model using UAV images were consistent with the manually segmented results (mean value and variance of <italic>Q</italic>_seg was 0.75 and 0.08), and the rice canopy coverage calculated by this method had a relatively high correlation with that calculated by manually segmentation (<italic>R</italic>²= 0.83, RMSE = 5.36%). The average rice plant height estimated by the canopy height model in each plot had a high correlation with the mean plant height measured in the field (<italic>R</italic>²= 0.81, RMSE = 9.81 cm), but it was underestimated as a whole. Based on the ground image, the panicle count results obtained by decision tree classification and morphological parameter filter had a relatively high correlation with the measured panicle number (<italic>R</italic>²= 0.83, RMSE = 10.99). Overall, combined with image analysis algorithm, using low-altitude UAV remote sensing technology to high-throughput and automatically estimate rice canopy coverage and plant height can achieve relatively high accuracy; using image from ground camera to accurately count the rice panicle number is of significant potentiality. The proposed pipeline in this research could be used to analyze nitrogen effect on rice growth status and evaluate nitrogen response of different rice cultivars, and it is of great significance for mining rice field phenotypic information and yield prediction.

An iterative morphological difference product wavelet for weak fault feature extraction in rolling bearing fault diagnosis

Weak fault feature extraction is of great significance to the fault diagnosis of rolling bearing. At the early stage of defects, fault features are usually weak and easily submerged in strong background noise, which makes feature information extremely difficult to be excavated. This paper proposes an iterative morphological difference product wavelet (MDPW) to address this issue. In this scheme, firstly, the morphological difference product filter (MDPF) is developed using the combination morphological filter-hat transform operator and difference operator. The MDPF is then incorporated into a morphological undecimated wavelet to construct the MDPW, which can achieve noise suppression and fault feature enhancement. Subsequently, the optimal iteration numbers that influence the performance of MDPW is determined using the fault severity indicator, which effectively extracts periodic impulse related to the failure of rolling bearing. Finally, the fault identification is inferred by the occurrence of fault defect frequencies in the MDPW spectrum with the optimal iteration numbers. The validity of the iterative MDPW is evaluated through numerical simulations and experiment cases. The analysis results demonstrate that the iterative MDPW has higher diagnosis accuracy than existing algorithms (e.g., adaptive single-scale morphological wavelet and weighted multi-scale morphological wavelet). This research provides a new perspective for improving the weak fault feature extraction of rolling bearing.

Automatic Video Surveillance System for Pedestrian Crossing Using Machine Learning

Nowadays accidents became very common in this world and the majority of accidents are from the crosswalks. The accidents are caused due to the absence of traffic lights to lead the traffic structure and people. Even though the traffic flow is huge by the traditional traffic accessories has a very particular warning capabilities which is incapable of to satisfy the safety requirements. Like a driver is unable to divine there will be pedestrian crossing the pathway forward, and pedestrians lack of traffic security alertness, which may easily cause of accidents. So, to overcome this issue the video surveillance helps the crippled persons to cross the pedestrians in a riskless way and it helps for independent vehicles. The work recommended an emerging technology to accommodate the crippled persons to cross the pathway just about the traffic signal and it also supports to someone who walks at a leisurely pace with the help of video surveillance. Here the image processing analysis algorithm takes a very crucial role to trace the motion of the object in the certain pedestrian crossing which can accommodate or serve the crippled persons or any senior citizens. They are numerous morphological filtering operations will improve the distinction of the motion of the person in the video. The proposed work improves more accuracy than the previous accuracy methods and also in addition of this we are implementing the crowd detection feature on the pedestrian, whenever the crowd was more in pedestrians.

Performance analysis of diabetic retinopathy using diverse image enhancement techniques

ABSTRACT Most of the retinal disease can be manifested by retinal fundus images. However, the fundus image quality is not adequate for diagnosing the retinal disease such as diabetic retinopathy (DR) due to colour distortion, low contrast, uneven illumination, and blurring. Therefore, there is a need for enhancing the images by applying various enhancement techniques. This paper implements diverse methods for image enhancement such as wiener filter, median filter, Contrast Stretching, Histogram Equalisation, Contrast adjustment, Morphological top hat filter, morphological bottom hat filter, Adaptive Histogram Equalisation, Contrast limited Adaptive histogram Equalisation (CLAHE) on retinal images. The performance has been evaluated on DRIVE and STARE dataset, which includes Peak Signal to noise Ratio, Structural similarity Index, Mean Square Error, Maximum difference, Normalised cross correlation, structural Content, Normalised absolute error, Average difference and Entropy of the images.

Automated deep learning approach for classification of malignant melanoma and benign skin lesions

Skin cancer becomes a significant health problem worldwide with an increasing incidence over the past decades. Due to the fine-grained differences in the appearance of skin lesions, it is very challenging to develop an automated system for benign-malignant classification through images. This paper proposes a novel automated Computer Aided Diagnosis (CAD) system for skin lesion classification with high classification performance using accuracy low computational complexity. A pre-processing step based on morphological filtering is employed for hair removal and artifacts removal. Skin lesions are segmented automatically using Grab-cut with minimal human interaction in HSV color space. Image processing techniques are investigated for an automatic implementation of the ABCD (asymmetry, border irregularity, color and dermoscopic patterns) rule to separate malignant melanoma from benign lesions. To classify skin lesions into benign or malignant, different pretrained convolutional neural networks (CNNs), including VGG-16, ResNet50, ResNetX, InceptionV3, and MobileNet are examined. The average 5-fold cross validation results show that ResNet50 architecture combined with Support Vector Machine (SVM) achieve the best performance. The results also show the effectiveness of data augmentation in both training and testing with achieving better performance than obtaining new images. The proposed diagnosis framework is applied to real clinical skin lesions, and the experimental results reveal the superior performance of the proposed framework over other recent techniques in terms of area under the ROC curve 99.52%, accuracy 99.87%, sensitivity 98.87%, precision 98.77%, F1-score 97.83%, and consumed time 3.2 s. This reveals that the proposed framework can be utilized to help medical practitioners in classifying different skin lesions.

EIT 40‐Electrodes Breast Cancer Detection and Screening

In this paper, we presented a new electrical impedance tomography (EIT) method for the early detection of breast cancer. We designed a breast phantom comprising three tumors of different sizes. We tested the reconstruction of the image of this breast using 16, 24, 32 and 40 electrodes. The electrodes were placed according to a static, adapted and dynamic mode. The quality of the reconstruction was carried out by the root mean square error (RMSE). The best result was obtained with 40‐electrodes system and dynamic mode. The image reconstruction was improved by a segmentation, an edge detection by Canny filter and a morphological closure filter. This system allowed the detection of a T1 tumor. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

Aerobic Exercise Fatigue Detection Based on Spatiotemporal Entropy and Label Technology

Excessive exercise can strengthen the body and make people happy, but it can also cause physical injury. To address this issue, this paper proposes the TFD-SE (Three-Frame Difference Spatiotemporal Entropy) algorithm and the LB (Label Propagation) algorithm, which are both based on SE (spatiotemporal entropy) and label technology. The TFD-SE algorithm calculates the difference image using the three-frame difference method, then calculates the SE of pixels in the difference image, and performs morphological filtering and threshold segmentation, allowing it to detect moving objects effectively. The significance value of unlabeled nodes in the image is calculated using the LB algorithm. In both qualitative and quantitative comparisons, the experimental results show that both algorithms outperform other classical algorithms in terms of detection performance.

Three-phase unbalanced voltage sag detection method based on morphological filtering

In dynamic voltage restorer (DVR), accurate and fast voltage sag detection method is the key. However, in the case of three-phase imbalance, the traditional voltage sag detection method still has shortcomings in real-time and accuracy. Therefore, a voltage sag detection method based on morphological filter is proposed. That is, based on the commonly used dq transformation method, the morphological filter is used to replace the Butterworth low-pass filter to suppress the second harmonic generated by the negative sequence component after dq transformation in the case of three-phase imbalance, reduce the system delay and improve the real-time and accuracy of the detection method. The morphological filter is designed, and the appropriate structural elements and morphological transformation algorithm are selected. The results show that the morphological filter has good filtering effect and small delay, and can better meet the detection requirements of DVR.

Melanoma segmentation using deep learning with test-time augmentations and conditional random fields

In a computer-aided diagnostic (CAD) system for skin lesion segmentation, variations in shape and size of the skin lesion makes the segmentation task more challenging. Lesion segmentation is an initial step in CAD schemes as it leads to low error rates in quantification of the structure, boundary, and scale of the skin lesion. Subjective clinical assessment of the skin lesion segmentation results provided by current state-of-the-art deep learning segmentation techniques does not offer the required results as per the inter-observer agreement of expert dermatologists. This study proposes a novel deep learning-based, fully automated approach to skin lesion segmentation, including sophisticated pre and postprocessing approaches. We use three deep learning models, including UNet, deep residual U-Net (ResUNet), and improved ResUNet (ResUNet++). The preprocessing phase combines morphological filters with an inpainting algorithm to eliminate unnecessary hair structures from the dermoscopic images. Finally, we used test time augmentation (TTA) and conditional random field (CRF) in the postprocessing stage to improve segmentation accuracy. The proposed method was trained and evaluated on ISIC-2016 and ISIC-2017 skin lesion datasets. It achieved an average Jaccard Index of 85.96% and 80.05% for ISIC-2016 and ISIC-2017 datasets, when trained individually. When trained on combined dataset (ISIC-2016 and ISIC-2017), the proposed method achieved an average Jaccard Index of 80.73% and 90.02% on ISIC-2017 and ISIC-2016 testing datasets. The proposed methodological framework can be used to design a fully automated computer-aided skin lesion diagnostic system due to its high scalability and robustness.

The phthalocyanine blue-green pigments devices intended for optical filters

The crystalline structure, optical, morphological and optical filter properties of magnesium phthalocyanine (MPc) blue–green pigments layers, grown onto glass substrate by the spin coating route, are studied. Besides the capacitance voltage (C-V) characteristics and resistance versus voltage of MPc organic heterojunctions are investigated showing a negative capacitance (NC) within forward bias voltage at higher frequency. Optical analysis demonstrates a top and valley in transmittance profile within visible range which leads to the occurrence of optical filter properties in organic devices. The (123)-oriented monoclinic lattice is revealed by X-ray patterns, optical transmission attains 67% inside Vis range, and the band gap is found to be of 2.9 eV. Nonlinear optical properties of Mg-Phthalocyanine are studied. Nanotips are revealed by atomic force microscope and a size around of 22.76 nm, and the RMS is of 10.5 nm are evaluated. Obtained parameters from C-V characteristics of MgPc/GaAs heterojunction are described and recorded as a result of applied frequency. Such extracted parameters are Nd ranging in 5.1-31.2x1019 cm−3 and Vbi of 0.18-0.39 V and ΦB of 0.36 –0.41 eV.

An efficient fuzzy deep learning approach to recognize 2D faces using FADF and ResNet-164 architecture

Deep learning using fuzzy is highly modular and more accurate. Adaptive Fuzzy Anisotropy diffusion filter (FADF) is used to remove noise from the image while preserving edges, lines and improve smoothing effects. By detecting edge and noise information through pre-edge detection using fuzzy contrast enhancement, post-edge detection using fuzzy morphological gradient filter and noise detection technique. Convolution Neural Network (CNN) ResNet-164 architecture is used for automatic feature extraction. The resultant feature vectors are classified using ANFIS deep learning. Top-1 error rate is reduced from 21.43% to 18.8%. Top-5 error rate is reduced to 2.68%. The proposed work results in high accuracy rate with low computation cost. The recognition rate of 99.18% and accuracy of 98.24% is achieved on standard dataset. Compared to the existing techniques the proposed work outperforms in all aspects. Experimental results provide better result than the existing techniques on FACES 94, Feret, Yale-B, CMU-PIE, JAFFE dataset and other state-of-art dataset.