Threshold Segmentation Algorithm Research Articles

Retinal Vessel Segmentation Based on Multiscale Matched Filtering

Retinal vascular function is complex, morphological structure varies from person to person, and is susceptible to vascular diseases and systemic vascular diseases. Its accurate segmentation is of great significance for disease diagnosis and identification. In this paper, a multi-scale matching filtering algorithm is proposed for the uneven size of retinal blood vessels. On the basis of the traditional singlescale Gaussian matching filter, multiscale Gaussian matched filters with two sizes are used to enhance grayscale images. Enhancement is performed, and the superimposed image is binarized using a twodimensional maximum entropy threshold segmentation algorithm. The algorithm is tested in the DRIVE database with sensitivity, specificity and accuracy of 0.803, 0.959, 0.981, respectively. Comparing with the traditional algorithm, the algorithm has high sensitivity, fast running speed and rich details of segmentation results.

Research on defect detection method for high-reflective-metal surface based on high dynamic range imaging

The high dynamic range imaging (HDRI) technology can realistically reproduce the scene and acquire more image information for analysis and processing. In this paper, a defect detection method based on HDRI is proposed to solve the visual inspection problem of industrial parts with high-reflective surface. Firstly, a fast calibration method is put forward to increase the calibration speed of the camera response function (CRF). Then, a novel algorithm named adaptive tone mapping based on color correction model in gradient domain is designed to adaptively adjust image chromaticity information. Finally, adaptive threshold segmentation and Haar-like feature extraction algorithm are used to detect surface defects in synthesized high dynamic range (HDR) images. The experimental results show that the HDR images processed by our algorithm can meet the detection requirements for high-reflective-metal surface. The number of sampling points required by our fast calibration algorithm of CRF is 1/10 times than that of the traditional algorithm, and the running time in the program is about 1/4 times than that of the traditional algorithm. What’s more, our improved local tone mapping algorithm is obviously improved in information entropy, contrast and average gradient.

Fracture characterization and permeability prediction by pore scale variables extracted from X-ray CT images of porous geomaterials

Pore scale variables (e.g., porosity, grain size) are important indexes to predict the hydraulic properties of porous geomaterials. X-ray images from ten types of intact sandstones and another type of sandstone samples subjected to triaxial compression are used to investigate the permeability and fracture characteristics. A novel double threshold segmentation algorithm is proposed to segment cracks, pores and grains, and pore scale variables are defined and extracted from these X-ray CT images to study the geometric characteristics of microstructures of porous geomaterials. Moreover, novel relations among these pore scale variables for permeability prediction are established, and the evolution process of cracks is investigated. The results indicate that the pore-scale permeability is prominently improved by cracks. In addition, excellent agreements are found between the measured and the estimated pore scale variables and permeability. The established correlations can be employed to effectively identify the hydraulic properties of porous geomaterials.

Fusing Multilevel Deep Features for Fabric Defect Detection Based NTV-RPCA

Fabric defect detection plays an important role in automated inspection and quality control in textile manufacturing. As the fabric images have complex and diverse textures and defects, traditional detection methods show a poor adaptability and low detection accuracy. Robust principal component analysis (RPCA) model that can be used to separate the image into object and background have proven applicable in fabric defect detection. However, how to represent texture feature of the fabric image more effectively is still problematic in this kind of method. In addition, the use of the traditional RPCA may result in low accuracy and more noises in sparse part. In this article, a novel fabric defect detection method based on multilevel deep features fusion and non-convex total variation regularized RPCA (NTV-RPCA) is proposed. Firstly, the image representation ability is well enhanced through multilevel deep features extracted by a convolutional neural network. Then, the non-convex total variation regularized RPCA is proposed in which total variation constraint significantly reduces the noises in sparse part and non-convex solution is more approximate to the authentic one. Next, multilevel saliency maps generated by the sparse matrixes are fused via RPCA to produce a more reliable detection result. Finally, the defect region is located by segmenting the fused saliency map via a threshold segmentation algorithm. Qualitative and quantitative experiments conducted on two public fabric image databases demonstrate that the proposed method improves the adaptability and detection accuracy comparing to the state-of-the-arts.

An improved Otsu threshold segmentation algorithm

An improved Otsu threshold segmentation algorithm

An improved Otsu threshold segmentation algorithm

Image segmentation is widely used as a fundamental step for various image processing applications. This paper focuses on improving the famous image thresholding method named Otsu's algorithm. Based on the fact that threshold acquired by Otsu's algorithm tends to be closer to the class with larger intraclass variance when the foreground and background have large intraclass variance difference, an improved strategy is proposed to adjust the threshold bias. We analysed the relationship between pixel greyscale value and the change of cumulative pixel number, and selected the ratio of pixel grey level value to a certain cumulative pixel number as the adjusted threshold. Experiments using typical testing images were set up to verify the proposed method both quantitatively and qualitatively. Two widely used metrics named misclassification error (ME) and dice similarity coefficient (DSC) were adopted for quantitative evaluation, and both quantitative and qualitative results indicated that the proposed algorithm could better segment the testing images and get competitive misclassification error and DSC values compared with Otsu's method and its improved versions proposed by Hu and Gong (2009) and Xu et al. (2011), and the time consumption of our method can be significantly reduced.

Image Segmentation for the Treatment Planning of Magnetic Resonance-Guided High-Intensity Focused Ultrasound (MRgHIFU) Therapy: A Parametric Study

In the present research work, image segmentation methods were studied to find internal parameters that provide an efficient identification of the regions of interest in Magnetic Resonance (MR) images used for the therapy planning of High-Intensity Focused Ultrasound (HIFU), a minimally invasive therapeutic method used for selective ablation of tissue. The involved image segmentation methods were threshold, level set and watershed segmentation algorithm with markers (WSAM), and they were applied to transverse and sagittal MR images obtained from an experimental setup of a murine experiment. A parametric study, involving segmentation tests with different values for the internal parameters, was carried out. The F-measure results from the parametric study were analyzed by region using Welch’s ANOVA followed by post hoc Games-Howell test to determine the most appropriate method for region identification. In transverse images, the threshold method had the best performance for the air region with a F-measure median of 0.9802 (0.9743–0.9847, interquartile range IQR 0.0104), the WSAM for the tissue, gel-pad, transducer and water region with a F-measure median of 0.9224 (0.8718–0.9468, IQR 0.075), 0.9553 (0.9496–0.9606, IQR 0.011), 0.9416 (0.9330–0.9540, IQR 0.021) and 0.9769 (0.9741–0.9803, IQR 0.0062), respectively. In sagittal images, threshold method had the best performance for the air region with a F-measure median of 0.9680 (0.9589–0.9735, IQR 0.0146), the WSAM for the tissue and gel-pad regions with a F-measure median of 0.9241 (0.8870–0.9426, IQR 0.0556) and 0.9553 (0.9472–0.9625, IQR 0.0153), respectively, and the Geodesic Active Contours (GAC) method for the transducer and water regions with a F-measure median of 0.9323 (0.9221–0.9402, IQR 0.0181) and 0.9681 (0.9627–0.9715, IQR 0.0088), respectively. The present research work integrates preliminary results to generate more efficient procedures of image segmentation for treatment planning of the MRgHIFU therapy. Future work will address the search of an automatic segmentation process, regardless of the experimental setup.

Classification of rice planthoppers based on shape descriptors

Here, classification of rice planthopper (RPH) based on shape descriptors was addressed to solve the low semantics problem of shape features in traditional RPH (mainly including the whiteback planthopper (Sogatella furcifera (Horvath)), the brown planthopper (Nilaparvata lugens (Stal)), and the small brown planthopper (Laodelphax striatellus (Fallen))) image classification research. Images of RPH were obtained from rice field by an automatic insect image acquisition device made by ourselves and insect images were divided into single images based on OTSU threshold segmentation algorithm. In terms of the images of RPH after segmentation, Fourier descriptors and Hu moments, which are from two aspects of contour curve and shape area, were extracted to describe shape features of RPH. Then, random forest (RF), an ensemble learning algorithm, was used as the classifier to distinguish RPH efficiently. The optimal number of trees and prediction variables of RF are chosen to be 150 and 4, respectively, by minimising the out-of-bag error. Experimental results show that classification accuracy of RPH based on shape descriptors reaches up to 93.93%. Therefore, it has been verified that the classification with the method presented here is accurate and semantic.

Adaptive Kaniadakis entropy thresholding segmentation algorithm based on particle swarm optimization

Kaniadakis entropy is a kind of generalized entropy based on the $$ \kappa $$ probability distribution, which has a good ability to deal with the distribution of long tail. The image thresholding algorithm based on Kaniadakis entropy can effectively segment images with long-tailed distribution histograms, such as nondestructive testing image. However, Kaniadakis entropy is a generalized information entropy with parameter. How to choose appropriate parameter $$ \kappa $$ is a problem to be solved. In this paper, we proposed an adaptive parameter selection Kaniadakis entropy thresholding algorithm. Based on a clustering effectiveness evaluation index, we transform the parameter selection problem into an optimization problem, then use particle swarm optimization search algorithm to optimize it and finally obtain the segmentation threshold under the optimal parameter. The presented algorithm can adaptively select parameters according to different images and obtain the optimal segmentation images. In order to show the effectiveness of the proposed method, the segmentation results are compared with several existing entropy-based thresholding algorithms. Experimental results both qualitatively and quantitatively demonstrate that the proposed method is effective.

An Adaptive Real-Time Detection Algorithm for Dim and Small Photoelectric GSO Debris.

Geosynchronous orbit (GSO) is the ideal orbit for communication, navigation, meteorology and other satellites, but the space of GSO is limited, and there are still a large number of space debris threatening the safety of spacecraft. Therefore, real-time detection of GSO debris is necessary to avoid collision accidents. Because radar is limited by transmitting power and operating distance, it is difficult to detect GSO debris, so photoelectric detection becomes the mainstream way to detect GSO debris. This paper presents an adaptive real-time detection algorithm for GSO debris in the charge coupled device (CCD) images. The main work is as follows: An image adaptive fast registration algorithm and an enhanced dilation difference algorithm are proposed. Combining with mathematical morphology, threshold segmentation and global nearest neighbor (GNN) multi-target tracking algorithm, the functions of image background suppression, registration, suspected target extraction and multi-target tracking are realized. The processing results of a large number of measured data show that the algorithm can detect dim geostationary earth orbit (GEO) and non-GEO debris in GSO belt stably and efficiently, and the processing speed meets the real-time requirements, with strong adaptive ability, and has high practical application value.

Image thresholding segmentation method based on minimum square rough entropy

Image thresholding based on rough entropy is an efficient image segmentation technique. The optimal thresholds of the existing exponential and logarithmic rough entropy thresholding segmentation algorithms have unclear physical meaning. In this paper, a new form of square rough entropy is defined to measure the roughness in an image, and the corresponding image thresholding segmentation algorithm is proposed. The novel square rough entropy has good properties and simple computation. In the proposed thresholding algorithm, the optimal threshold is at the boundary between the object and the background of an image. This process is consistent with the expectation of image bi-level thresholding segmentation. To effectively granulate the image, a granule size selection method based on the homogeneity histogram is proposed, which is helpful in taking care of small objects and local variations of the images. The proposed thresholding method is evaluated by comparison with other three existing rough entropy based thresholding methods and three state-of-the-art image thresholding methods both qualitatively and quantitatively, using the natural images, the non-destructive testing images and an infrared video sequence from the OTCBVS Benchmark Data set. The comparison confirms the effectiveness of the proposed algorithm that not only is simple in form and clear in meaning, but also has good segmentation effect.

Analysis and detection of decayed blueberry by low field nuclear magnetic resonance and imaging

Decay is a major cause of quality loss and food safety concerns for blueberry fruit. Low-field nuclear magnetic resonance (LF-NMR) has been used to analyze and detect decayed blueberry fruit. According to NMR relaxation analysis, transverse relaxation time (T21, T22, T23) of decayed fruit increased to different degrees, while signal amplitude (A21, A22, A23) decreased to different degrees. Six relaxation features were input into BPNN model to identify the decay classes of fruit. The identification accuracy of training set was 86.7%, and that of the validation set was 90%. According to magnetic resonance imaging (MRI), twice threshold segmentation algorithm was proposed to segment the decayed region with dark color. A total of 15 features were extracted from gray histogram, gray level co-occurrence matrix (GLCM) and gray level-gradient co-occurrence matrix (GGCM) of the image for correlation analysis. Five features were then input into a back propagation neural network (BPNN) model. The identification accuracy of training set was 92.2%, and that of the validation set was 83.3%. Finally, 11 variables, including transverse relaxation features and image features, were input into the BPNN model. The identification accuracy of training set was 98.8%, and that of the validation set was 94.2%, which showed the highest identification accuracy. The results showed that LF-NMR and MRI are suitable for analyzing and detecting the decay disease of fruit, which provides a theoretical basis for nondestructive detection of fruit disease.

Centroid extraction algorithm based on grey-gradient for autonomous star sensor

The noise and complicated working environment experienced by the star sensor during spacecraft applications both have negative effects on the star image quality, which consequently degrades centroid extraction precision. This study proposes a novel centroid extraction algorithm to improve the precision and anti-interference performance. First, the block adaptive threshold segmentation (BATS) algorithm is proposed as a preliminary method for reducing star image noise. Then, to overcome problems caused by overlapping star spots and to provide dependable seed points for the local region growth (LRG) method, a grey-gradient-based greyscale cross-projection method (GGCPM) is proposed. Finally, LRG is combined with Gaussian surface fitting method to accurately extract the centroid since the edge of star spot could be maintained to further improve centroid extraction accuracy. Numerical simulation results demonstrate that the proposed algorithm has higher precision and better anti-interference performance than several traditional centroid extraction algorithms, and an actual star image application showed the algorithm is reliable and stable when applied in an actual environment.

Research and analysis of threshold segmentation algorithms in image processing

Image segmentation is one of the most difficult and important tasks in digital image processing. The accuracy and effect of segmentation determine the final success or failure of the calculation and analysis process. Therefore, analysis and improvement on basic image segmentation algorithms should be paid attention to in a quite wide range of applications. This paper mainly introduces and studies the principle and characteristics of threshold segmentation algorithm, and analyses the application scenarios of global threshold segmentation and adaptive local threshold segmentation related algorithms, which has certain reference significance for digital image processing related research.

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.

Combining SUN-based visual attention model and saliency contour detection algorithm for apple image segmentation

Accurate segmentation of apple fruit under natural illumination conditions provides benefits for growers to plan relevant applications of nutrients and pesticides. It also plays an important role for monitoring the growth status of the fruit. However, the segmentation of apples throughout various growth stages had only achieved a limited success so far due to the color changes of apple fruit as it matures as well as occlusion and the non-uniform background of apple images acquired in an orchard environment. To achieve the segmentation of apples with different colors and with various illumination conditions for the whole growth stage, a segmentation method independent of color was investigated. Features, including saliency and contour of the image, were combined in this algorithm to remove background and extract apples. Saliency using natural statistics (SUN) visual attention model was used for background removal and it was combined with threshold segmentation algorithm to extract salient binary region of apple images. The centroids of the obtained salient binary region were then extracted as initial seed points. Image sharpening, globalized probability of boundary-oriented watershed transform-ultrametric contour map (gPb-OWT-UCM) and Otsu algorithms were applied to detect saliency contours of images. With the built seed points and extracted saliency contours, a region growing algorithm was performed to accurately segment apples by retaining as many fruit pixels and removing as many background pixels as possible. A total of 556 apple images captured in natural conditions were used to evaluate the effectiveness of the proposed method. An average segmentation error (SE), false positive rate (FPR), false negative rate (FNR) and overlap Index (OI) of 8.4, 0.8, 7.5 and 90.5% respectively, were achieved and the performance of the proposed method outperformed other six methods in comparison. The method developed in this study can provide a more effective way to segment apples with green, red, and partially red colors without changing any features and parameters and therefore it is also applicable for monitoring the growth status of apples.

Mature pomegranate recognition methods in natural environments using machine vision

ABSTRACT: The use of machine vision to recognize mature pomegranates in natural environments is of major significance in improving the applicability and work efficiency of picking robots. By analyzing the color characteristics of color images of mature pomegranates under different illumination conditions, the feasibility of the YCbCr color model for pomegranate image recognition under different illumination conditions was proven. First, the Cr component map of pomegranate image is selected and then the pomegranate fruit is segmented by the kernel fuzzy C-means clustering algorithm to obtain the pomegranate image. Contrast experiments of pomegranate image segmentation under different illumination conditions were then performed using the proposed kernel fuzzy C-means clustering algorithm, the fuzzy C-means clustering algorithm, the Otsu algorithm and the threshold segmentation algorithm. Results of the experiments verified the effectiveness and superiority of the proposed algorithm.

A novel approach-foot tumour detection algorithm using watershed and thresholding based segmentation.

Detection of foot tumour has always been a little tedious as it is generally confounded with other diseases of the foot. Hence, the method of detecting foot tumour is of utmost importance. Though image processing techniques have replaced the original biopsy method, these classical methods are still attributed to a few drawbacks. In this paper we have proposed an efficient methodology which extracts the image and converts an input color image to YcBcRc components and each component of YcBcRc component are extracted separately to detect the exact size and location of tumour region in the foot using threshold and watershed segmentation algorithms by applying morphological operators. After the scanned MRI is enhanced using segmentation, the tumour can be detected effectively.

Rapid Detection of Rice Disease Based on FCM-KM and Faster R-CNN Fusion

In this paper, a method for detecting rapid rice disease based on FCM-KM and Faster R-CNN fusion is proposed to address various problems with the rice disease images, such as noise, blurred image edge, large background interference and low detection accuracy. Firstly, the method uses a two-dimensional filtering mask combined with a weighted multilevel median filter (2DFM-AMMF) for noise reduction, and uses a faster two-dimensional Otsu threshold segmentation algorithm (Faster 2D-Otsu) to reduce the interference of complex background with the detection of target blade in the image. Then the dynamic population firefly algorithm based on the chaos theory as well as the maximum and minimum distance algorithm is applied for optimization of the K-Means clustering algorithm (FCM-KM) to determine the optimal clustering class k value while addressing the tendency of the algorithm to fall into the local optimum problem. Combined with the R-CNN algorithm for the identification of rice diseases, FCM-KM analysis is conducted to determine the different sizes of the Faster R-CNN target frame. As revealed by the application results of 3010 images, the accuracy and time required for detection of rice blast, bacterial blight and blight were 96.71%/0.65s, 97.53%/0.82s and 98.26%/0.53s, respectively, indicating clearly that the method is more capable of detecting rice diseases and improving the identification accuracy of Faster R-CNN algorithm, while reducing the time required for identification.

Human Lesion Detection Method Based on Image Information and Brain Signal

The brain is the largest and most complex structure in the central nervous system. It dominates all activities in the body, and the lesions in the human body are also reflected in the brain signal. In this paper, the image method is used to assist the brain signal to detect the human lesion. Due to the particularity of medical images, there is no common segmentation method for any medical image, and there is no objective standard to judge whether the segmentation is effective. Medical image segmentation technology is still a bottleneck restricting the development and the application of other related technologies in medical image processing. Based on the above reasons, this paper proposes an improved region growing algorithm based on the fuzzy theory and region growing algorithm. The algorithm is used to segment the medical images of the liver and chest X-ray of different human organs. The improved algorithm uses a threshold segmentation algorithm to assist in the automatic selection of seed points and improves the region growing rules, then morphological post-processing is used to improve the segmentation effect. The experimental results show that the improved region growing algorithm has better segmentation effect under two different organs, which proves that the algorithm has certain applicability, and its accuracy and segmentation quality are better than the traditional region growing algorithm. This algorithm combines the advantages of the threshold method and traditional region growing method. It is feasible in algorithm and has certain application value.