Global Thresholding Algorithm Research Articles

Evaluation of Flaw Detection Algorithm Using Simulated X-Ray Computed Tomography of Ground Truth Data

Abstract A framework to generate simulated X-ray computed tomography (XCT) data of ground truth (denoted here as “GT”) flaws was developed for the evaluation of flaw detection algorithms using image comparison metrics. The flaws mimic some of those found in additively manufactured parts. The simulated flaw structure gave a GT data set with which to quantitatively evaluate, by calculating exact errors, the results of flaw detection algorithms applied to simulated XCT images. The simulated data avoided time-consuming manual voxel labeling steps needed for many physical data sets to generate GT images. The voxelated pore meshes that exactly match GT images were used in this study as opposed to using continuum pore meshes. The voxelated pore mesh approach avoids approximation error that occurs when converting continuum pore meshes to voxelated GT images. Spherical pores of varying sizes were randomly distributed near the surface and interior of a cylindrical part. XCT simulation was carried out on the structure at three different signal-to-noise levels by changing the number of frames integrated for each projection. Two different local thresholding algorithms (a commercial code and the Bernsen method) and a global thresholding algorithm (Otsu) were used to segment images using varying sets of algorithm parameters. The segmentation results were evaluated with various image evaluation metrics, which showed different behaviors for the three algorithms regarding “closeness” to the GT data. An approach to optimize the thresholding parameters was demonstrated for the commercial flaw detection algorithm based on semantic evaluation metrics. A framework to evaluate pore sizing error and binary probability of detection was further demonstrated to compare the optimization results.

Showthrough and Strikethrough print defect detection using histogram equalization based computer vision method

This paper presents a comparatively simple approach for showthrough and strikethrough print defect detection using computer vision method. Showthrough and strikethrough are common printing problem and are typically functions of a paper’s opacity. Under normal lighting condition the visibility of printing on the reverse side of printed paper is termed as showthrough whereas the penetration of ink to the other side is termed as strikethrough. Moreover the intensity of showthrough pixel is extremely low thus it is difficult to identify the showthrough pixel from the printed area. On the other hand strikethrough is the result of penetration of ink through paper and depends on the absorbent nature of paper. Comparatively the intensity of the strikethrough pixel is higher than that of the showthrough but due to similar intensity of the ink of the printed pixel and strikethrough pixel, both overlapped with each other in the foreground of the image. These print defects can degrade the image quality as well as print production. In this study, the detection of these two print defects achieved using histogram equalization technique, to enhance the contrast between foreground and back ground pixels. A global thresholding algorithm was applied on a his-togram equalized image to segment the printed area from the background of the image. Pixels in the background which are considered as showthrough and strike through pixels are identified by image subtraction. The pictorial representations of the results show the remarkable potential of the proposed technique which can be possible alternative of present subjective measures of showthrough and strikethrough.

Automated and ImageJ thresholding algorithm-based analysis of macular vessel density in diabetic patients.

Purpose:To assess the macular vessel density (VD) on optical coherence tomography angiography (OCT-A) using proprietary software (automated) and image processing software (manual) in diabetic patients.Methods:In a retrospective study, OCT-A images (Triton, TOPCON Inc.) of type 2 diabetics presenting to a tertiary eye care center in North India between January 2018 and December 2019 with or without nonproliferative diabetic retinopathy (NPDR) and with no macular edema were analyzed. Macular images of size 3 × 3 mm were binarized with global thresholding algorithms (ImageJ software). Outcome measures were superficial capillary plexus VD (SCP-VD, automated and manual), deep capillary plexus VD (DCP-VD, manual), and correlation between automated and manual SCP-VD.Results:OCT-A images of 89 eyes (55 patients) were analyzed: no diabetic retinopathy (NoDR): 29 eyes, mild NPDR: 29 eyes, and moderate NPDR: 31 eyes. Automated SCP-VD did not differ between NoDR and mild NPDR (P = 0.69), but differed between NoDR and moderate NPDR (P = 0.014) and between mild and moderate NPDR (P = 0.033). Manual SCP-VD (Huang and Otsu methods) did not differ between the groups. Manual DCP-VD differed between NoDR and mild NPDR and between NoDR and moderate NPDR, but not between mild and moderate NPDR with both Huang (P = 0.024, 0.003, and 0.51, respectively) and Otsu (P = 0.021, 0.006, and 0.43, respectively) methods. Automated SCP-VD correlated moderately with manual SCP-VD using Huang method (r = 0.51, P < 0.001) with a mean difference of −0.01% (agreement limits from −6.60% to +6.57%).Conclusion:DCP-VD differs consistently between NoDR and NPDR with image processing, while SCP-VD shows variable results. Different thresholding algorithms provide different results, and there is a need to establish consensus on the most suited algorithm.

A Study of Left Ventricular (LV) Segmentation on Cardiac Cine-MR Images

Left ventricular segmentation from cardiac images has high impact to have early diagnosis of various cardiovascular disorders. However, it is really a challenging task to segment left ventricular images from magnetic resonance image (MRI). In this paper, we explore several state-of-the-art segmentation algorithms applied on left ventricular (LV) segmentation on cardiac cine-MR images. Both adaptive and global thresholding algorithms along with region-based segmentation algorithm have been explored. Edge-based segmentation is disregard due to the absence of edge information in the employed dataset. For evaluation, we explored a benchmark dataset that was used for the MICCAI 3D segmentation challenge. We found that the cardiac MRI global thresholding has proved to be much efficient and robust than the adaptive thresholding. We achieved more than 92% accuracy for global thresholding, whereas, about 78% accuracy for the adaptive thresholding approach. The use of entropy or histogram to characterize segmentation in place of the intensity value of the pixel has a vital effect on segmentation efficiency. It is evident that the intensity information is corrupted by acquisition procedure, as well as the structure of organs. Due to the lack of boundary information in cardiac cine-MRI, clustering and region-based segmentation have produced more than 93% segmentation accuracy. For the case of soft clustering, the increased accuracy is found as 96%. However, more explorations are required, specially based on deep learning approaches on very large datasets.

Trainable segmentation for transmission electron microscope images of inorganic nanoparticles.

We present a trainable segmentation method implemented within the python package ParticleSpy. The method takes user labelled pixels, which are used to train a classifier and segment images of inorganic nanoparticles from transmission electron microscope images. This implementation is based on the trainable Waikato Environment for Knowledge Analysis (WEKA) segmentation, but is written in python, allowing a large degree of flexibility and meaning it can be easily expanded using other python packages. We find that trainable segmentation offers better accuracy than global or local thresholding methods and requires as few as 100 user-labelled pixels to produce an accurate segmentation. Trainable segmentation presents a balance of accuracy and training time between global/local thresholding and neural networks, when used on transmission electron microscope images of nanoparticles. We also quantitatively investigate the effectiveness of the components of trainable segmentation, its filter kernels and classifiers, in order to demonstrate the use cases for the different filter kernels in ParticleSpy and the most accurate classifiers for different data types. A set of filter kernels is identified that are effective in distinguishing particles from background but that retain dissimilar features. In terms of classifiers, we find that different classifiers perform optimally for different image contrast; specifically, a random forest classifier performs best for high-contrast ADF images, but that QDA and Gaussian Naïve Bayes classifiers perform better for low-contrast TEM images.

RETRACTED ARTICLE: Monitoring of mountain ecological environment based on Bayesian estimation and testing of motor memory function in mice

In this paper, the Bayesian estimation method is used to estimate the coefficients of wavelet transform to reconstruct the desired signal and reduce the noise. Using SNR and RMS error as noise reduction estimation indexes, we compare Bayesian estimation algorithm and global threshold algorithm, which improves the shortcomings of threshold algorithm and adaptive threshold algorithm. According to Bayesian estimation, according to the geological vulnerability characteristics of the mountain ecological environment, the geological change and its promotion mechanism of the mountain environment are selected. The paper analyzes the influence of LUCC remote sensing monitoring and dynamic changes, such as rainfall erosion and topography, and the influence of individual natural management factors, and analyzes the propulsion effect of nearby residential areas. It combines natural management with artificial management, and discusses deeply in various aspects, including the assessment and analysis of environmental geological vulnerability, and obtains relevant research results. This paper studies the function of motor memory in mice experimental model, and provides experimental basis for PD treatment in GTP. The methods were 25-30g male mice were randomly divided into normal group, control group, low dose GTP group, and high dose GTP group. After modeling mice in the control group, low dose GTP group, and high dose GTP group, climbing bar, and other experiments were carried out to detect. By studying Bayesian estimation and mountain ecological environment monitoring, it was applied to the detection of motor memory function in mice, which promoted the development of motor memory function.

Comparison of thresholding methods for shoreline extraction from Sentinel-2 and Landsat-8 imagery: Extreme Lake Salda, track of Mars on Earth

Lake Salda's extreme environment is geologically similar to Jezero Crater paleolake on Mars due to the formation of stromatolites and its extremely alkaline and cold water. It is critical to accurately determine the shoreline in the littoral zone where stromatolite formation occurs, in alkaline clean lakes like Salda, which contain traces of life on Mars, and in monitoring the change that occurs with climate and anthropogenic effect. The performance of global automatic thresholding algorithms on shoreline determination from NDWI and mNDWI water indices is compared in this study using Sentinel-2 and Landsat-8 images atmospherically corrected by different algorithms. Satellite images data acquired on August 12, 2020 for Sentinel-2 and August 11, 2020 for Landsat-8 on Lake Salda were used to determining the shoreline. The shoreline data measured in situ concurrently with the Sentinel-2 satellite acquisition was used as reference data. In the accuracy analysis, ground control points created inside and outside the lake at a distance of 1 pixel and 0.5 pixel to the reference shoreline for each satellite image were used. The performance of the optimal threshold values determined by each thresholding algorithm in the water index images was assessed using Kappa coefficient, Overall Accuracy (OA), %OA of Inside (%OAinside) and %OA of Outside (%OAoutside) statistics metrics. The optimal threshold values vary depending on the image and the atmospheric correction algorithm applied to the image. The NDWI index produces more accurate results in both Sentinel-2 and Landsat-8 satellite images. While atmospheric correction algorithms do not affect the results in Landsat-8 images, the Sen2Cor algorithm outperforms iCOR in Sentinel-2 images. For thresholding algorithms to be used in different water index and satellite images, Intermode, Isodata, IJ_Isodata, Minimum and Otsu algorithms in Landsat-8_LaSRC_NDWI and Landsat-8_iCOR_NDWI images, and Intermode, Minimum and Huang algorithms in Sentinel-2_Sen2Cor_NDWI images produce the best results. Because the Minimum algorithm causes significant gaps in the lake surface, the Huang and Intermode algorithms should be used for Sentinel-2_Sen2Cor_NDWI images. The 0 (zero) threshold value in the water indices images has a high accuracy only in the NDWI water indices generated from the Landsat-8 image.

Determination of the shear failure areas of rock joints using a laser scanning technique and artificial intelligence algorithms

Two artificial intelligence models-backpropagation neural network (BPNN) and support vector machines (SVM)-were created to investigate the effects of mesostructure characteristics on the shear mechanical behaviors of rock joints. This was achieved through learning training samples for the evaluation of the five basic geometrical parameters sensitivity (i.e., slope angle, horizontal orientation, elevation difference, curvature, and aperture distribution), and for determination of the shear failure regions when rock joints were subjected to low-normal and shear loads. First, the digital elevation models (DEMs) of rock joints were produced through point clouds collected using a laser scanning system. Five geometrical parameters were specified as the inputs for the artificial intelligence models, and an approach was developed to calculate the 3D aperture distribution using a point cloud registration algorithm. Shear failure regions were considered as the outputs, which were extracted from images taken after direct shear testing via the global threshold algorithm. Secondly, BPNN and SVM models were employed in order to establish relationships between the geometrical parameters and shear failure areas by machine learning on training samples. Thirdly, the information value (IV) algorithm was used to verify the two trained models. Results showed that the BPNN and SVM models made acceptable determinations of the shear failure areas, which corresponded to the real situation. The predictions from the BPNN and SVM models were more accurate than those from the IV algorithm. Furthermore, shear failure regions depended primarily on the aperture distribution of rock joints during the shearing process, followed by horizontal orientation, elevation difference, and then the slope angle. The effect of parameter curvature on shear failure was less than the other four parameters.

Fully Parallel Implementation of Otsu Automatic Image Thresholding Algorithm on FPGA

This work proposes a high-throughput implementation of the Otsu automatic image thresholding algorithm on Field Programmable Gate Array (FPGA), aiming to process high-resolution images in real-time. The Otsu method is a widely used global thresholding algorithm to define an optimal threshold between two classes. However, this technique has a high computational cost, making it difficult to use in real-time applications. Thus, this paper proposes a hardware design exploiting parallelization to optimize the system’s processing time. The implementation details and an analysis of the synthesis results concerning the hardware area occupation, throughput, and dynamic power consumption, are presented. Results have shown that the proposed hardware achieved a high speedup compared to similar works in the literature.

Spectrum decomposition in Gaussian scale space for uneven illumination image binarization.

Although most images in industrial applications have fewer targets and simple image backgrounds, binarization is still a challenging task, and the corresponding results are usually unsatisfactory because of uneven illumination interference. In order to efficiently threshold images with nonuniform illumination, this paper proposes an efficient global binarization algorithm that estimates the inhomogeneous background surface of the original image constructed from the first k leading principal components in the Gaussian scale space (GSS). Then, we use the difference operator to extract the distinct foreground of the original image in which the interference of uneven illumination is effectively eliminated. Finally, the image can be effortlessly binarized by an existing global thresholding algorithm such as the Otsu method. In order to qualitatively and quantitatively verify the segmentation performance of the presented scheme, experiments were performed on a dataset collected from a nonuniform illumination environment. Compared with classical binarization methods, in some metrics, the experimental results demonstrate the effectiveness of the introduced algorithm in providing promising binarization outcomes and low computational costs.

Retracted] Performance Analysis of Otsu‐Based Thresholding Algorithms: A Comparative Study

Image thresholding is a widely used technology for a lot of computer vision applications, and among various global thresholding algorithms, Otsu‐based approaches are very popular due to their simplicity and effectiveness. While the usage of Otsu‐based thresholding methods is well discussed, the performance analyses of these methods are rather limited. In this paper, we first review nine Otsu‐based approaches and categorize them based on their objective functions, preprocessing, and postprocessing strategies. Second, we conduct several experiments to analyze the model characteristics using different scene parameters both on synthetic images and real‐world cell images. We put more attention to examine the variance of foreground object and the effect of the distance between mean values of foreground and background. Third, we explore the robustness of algorithms by introducing two typical kinds of noises under different intensities and compare the running time of each method. Experimental results show that NVE, WOV, and Xing’s methods are more robust to the distance of mean values of foreground and background. The large foreground variance will cause a larger threshold value. Experiments on cell images show that foreground miss detection becomes serious when the intensities of foreground pixels change drastically. We conclude that almost all algorithms are significantly affected by Salt&amp;Pepper and Gaussian noises. Interestingly, we find that ME increases almost linearly with the intensity of Salt&amp;Pepper noise. In terms of algorithms’ time cost, methods with no preprocessing and postprocessing steps have more advantages. All these findings can serve as a guideline for image thresholding when using Otsu‐based thresholding approaches.

Skid resistance evaluation of pervious pavement mixtures using XRCT-based modelling

Abstract Pervious pavements can help mitigate climate change effects while improving transportation safety by improving wet pavement friction and reducing splash and spray. Prevailing pervious pavement mix design procedures adopt laboratory-scale friction experiments which cannot capture field wet tire-pavement friction performance. To bridge this gap, this paper presents the application of a newly developed discharge-based thresholding algorithm for wet pervious pavement skid resistance estimation. In particular, x-ray computed tomography (XRCT) scanning, digital image processing (DIP) algorithms and finite-element modelling of wet tire-pavement interaction are adopted to bridge laboratory experiments and field performance. Our developed algorithm is found to be superior in performance when compared against other existing global thresholding algorithms in the literature. It was found from the case study that our developed framework is capable of predicting field skid resistance of various pervious pavement mixtures at the design stage, thereby aiding in the selection of friction-efficient pervious pavement mixtures.

Sugeno integral generalization applied to improve adaptive image binarization

Classic adaptive binarization methodologies threshold pixels intensity with respect to adjacent pixels exploiting integral images. In turn, integral images are generally computed optimally by using the summed-area-table algorithm (SAT). This document presents a new adaptive binarization technique based on fuzzy integral images. Which, in turn, this technique is supported by an efficient design of a modified SAT for generalized Sugeno fuzzy integrals. We define this methodology as FLAT (Fuzzy Local Adaptive Thresholding). Experimental results show that the proposed methodology produced a better image quality thresholding than well-known global and local thresholding algorithms. We proposed new generalizations of different fuzzy integrals to improve existing results and reaching an accuracy ≈0.94 on a wide dataset. Moreover, due to high performances, these new generalized Sugeno fuzzy integrals created ad hoc for adaptive binarization, can be used as tools for grayscale processing and more complex real-time thresholding applications.

Apple fruit sorting using novel thresholding and area calculation algorithms

The speed and capacity (kg/hr) for sorting apples are two important factors that have to be taken care of post-harvest. To increase the speed and capacity of apples being sorted, a new algorithm for segmenting the region of interest from an apple fruit image and to find the percentage of defective area after sorting process has been proposed in this paper. First, the region of interest from the acquired color image of an apple fruit was segmented using the proposed global thresholding algorithm GTA. Features were extracted from the segmented image using the coefficients obtained by applying wavelet transformation using the Haar filter. Further, the Naive Bayesian classifier was applied to sort apples as defective and sound. Second, the defective area of the defective sample was segmented using the K-means algorithm and median filter. The defective area in percentage was calculated using the proposed diseased area calculation algorithm DACA to decide about the acceptance of that defective sample. The performance of the proposed GTA was compared with Otsu’s and Kapur’s thresholding algorithms. With less segmentation time, it was found that the Naive Bayesian classification using GTA gave a 96.67% accuracy rate than Otsu’s with 65% and Kapur’s with 93.33% accuracy rate. The execution time of the proposed DACA was less when compared to the bwarea function in MATLAB.

Quantification of diabetic macular ischemia using novel three-dimensional optical coherence tomography angiography metrics.

We applied three-dimensional (3D) analysis to optical coherence tomography angiography (OCTA) to measure macular ischemia in eyes affected by non-proliferative diabetic retinopathy (DR). A previously validated algorithm was applied to OCTA data in order to obtain 3D visualization of the retinal vasculature. Successively, a global thresholding algorithm was applied and two novel quantitative metrics were introduced: 3D vascular volume and 3D perfusion density. Two-dimensional (2D) OCTA metrics were also obtained with different binarization thresholds for comparison. Of the 30 patients included, 15 were diagnosed with DR and 15 were controls. The 3D vascular volume and 3D perfusion density were reduced in DR eyes (P < .0001). The 2D variables also significantly differ between groups. The 3D perfusion density had the highest area under the receiver operating characteristic curve (0.964) among tested variables. Assessing quantitative perfusion using 3D analysis is reliable and promising, and with an elevated diagnostic efficacy in identifying DR eyes.

The artificial neural network to predict chlorophyll content of cassava (Manihot esculenta) leaf

Artificial neural network (ANN) based prediction system was presented for predicting the leaf population chlorophyll content from cassava leaf images. As the training of this prediction system relied heavily on how well those leaf green pixels were separated from background noises in cassava leaf images, a global thresholding algorithm and an omnidirectional scan noise filtering coupled with the hue histogram statistic method were designed for leaf green pixel extraction. With the obtained of leaf green pixels, the system training was carried out by applying a back-propagation algorithm. The system was tested to predict the chlorophyll content from the cassava leaf images. The purpose of this research was to find the relationship between the color index Red, Green, Blue (RGB); Hue; Saturation HSV; Value; Saturation HSL and Lightness to chlorophyll, and to find the appropriate form of ANN to predict the highest chlorophyll source in cassava leaf based on digital images. The results showed highest positive regression occurred in the saturation HSL index against the total chlorophyll of cassava leaf as much as 78.6%. The best model produced using ANN methods in predicting total chlorophyll is a network model with 8 inputs, 9 hidden layers and one output layer, in the proportion of training data 75% testing data 25% have value result the smallest MSE testing is 0.092 with regression testing of 0.847. Network model can read the highest source of chlorophyll on cassava leaf with the value of 84.68%.

Artistic multi-script identification at character level with extreme learning machine

In this work, a novel problem, namely artistic multi-script identification at character level has been addressed. Two types of documents: real/ natural and synthetic have been used for dataset preparation. After binarizing using Otsu’s global thresholding algorithm, a semi-automatic segmentation technique has been applied for character separation. Some well-known texture based features have been considered from the segmented images and further, they have been converted into lower dimensional space by applying principal component analysis. Those final feature set are classified using an Extreme Learning based classifier and performance are compared with traditional machine learning techniques and other features. Observing the inherent complexity of the multi-script character level datasets, an encouraging outcome has been obtained.

Binarizing water sensitive papers – how to assess the coverage area properly?

Water sensitive paper is a convenient tool used to assess the coverage of an area with liquid in many various applications. It is useful for visualization and evaluation of potential spray deposits on plants. This article attempts to guide readers in the quantification of the coverage area of the water sensitive paper with liquid. It summarises a comparison of 15 global thresholding and binarization algorithms for their ability to quantify the coverage area accurately. We used 40 water sensitive papers obtained from two field experiments, in which the quality of spraying potatoes and rape was tested. Obtained results show that when binarizing water sensitive papers, careful choice of thresholding method is necessary, as differences in the results obtained with the use of various thresholding algorithms were observed and they were quite substantial in some cases. In relation to expert assessment, the algorithms proposed by Huang and Wang (1995), Ridler and Calvard (1978) and Otsu (1979), provided the best results.

An improved binarization algorithm of wood image defect segmentation based on non-uniform background

In this study, an image binarization optimization algorithm, based on local threshold algorithms, is proposed because global and traditional local threshold segmentation algorithms cannot effectively address the problems of non-uniform backgrounds of wood defect images. The proposed algorithm calculates the threshold by the mean, standard deviation and the extreme value of the window. The results indicate that this modified algorithm enhances the image segmentation for wood defect images on a complex background, which is much superior to the global threshold algorithm and the Bernsen algorithm, and slightly better than the Niblack algorithm and Sauvola algorithm. Compared with similar models, the algorithm proposed in this paper has higher segmentation accuracy, as high as 92.6% for wood defect images with a complex background.

Local thresholding of degraded or unevenly illuminated documents using fuzzy inclusion and entropy measures

There are applications in which the content of a scanned document needs to be recognized or improved. We often achieve this by converting our input into a binary image and this is in fact the first step in many document analysis systems or optical character recognition (OCR) processes. In cases where our input is degraded or has a non-uniform illumination, global thresholding algorithms fail to deliver adequate results. For this reason, we have to use some local thresholding techniques which binarize each pixel based on the grayscale information of its adjoining pixels. In this paper, we present a local thresholding method based on specific fuzzy inclusion and entropy measures which we introduced in some of our previous work. We use these indicators to measure specific attributes of the neighborhood of a pixel and then, based on these values, an appropriate threshold is calculated. We don’t use the histogram of the image or any statistical measures and contrast parameters depending on the input. It is an open, automated and adaptable procedure and in this presentation we see some implementations of a more general algorithm along with some specific results. Our main domain of experimentation consists of texts containing lighting “irregularities” but some remarks regarding further generalization are being made as well. We also comment on other potential of these measures and the prospect of being connected with other studies that already use fuzzy inclusion and entropy measures.