Region-growing Method Research Articles

Sharp Feature Detection as a Useful Tool in Smart Manufacturing

Industry 4.0 comprises a wide spectrum of developmental processes within the management of manufacturing and chain production. Presently, there is a huge effort to automate manufacturing and have automatic control of the production. This intention leads to the increased need for high-quality methods for digitization and object reconstruction, especially in the area of reverse engineering. Commonly used scanning software based on well-known algorithms can correctly process smooth objects. Nevertheless, they are usually not applicable for complex-shaped models with sharp features. The number of the points on the edges is extremely limited due to the principle of laser scanning and sometimes also low scanning resolution. Therefore, a correct edge reconstruction problem occurs. The same problem appears in many other laser scanning applications, i.e., in the representation of the buildings from airborne laser scans for 3D city models. We focus on a method for preservation and reconstruction of sharp features. We provide a detailed description of all three key steps: point cloud segmentation, edge detection, and correct B-spline edge representation. The feature detection algorithm is based on the conventional region-growing method and we derive the optimal input value of curvature threshold using logarithmic least square regression. Subsequent edge representation stands on the iterative algorithm of B-spline approximation where we compute the weighted asymmetric error using the golden ratio. The series of examples indicates that our method gives better or comparable results to other methods.

Automatic method for segmenting leaves by combining 2D and 3D image-processing techniques.

In this study, a method to automatically segment plant leaves from three-dimensional (3D) images using structure from motion is proposed. First, leaves in the 3D images are roughly segmented using a region-growing method in which near points with distances less than 0.2 cm are assigned to the same group. By repeating this process, the leaves not touching each other can be segmented. Then, each segmented leaf is projected onto two-dimensional (2D) images, and the watershed algorithm is executed. This process successfully segments overlapping leaves.

An optimised pixel-based classification approach for automatic white blood cells segmentation

The pixel-based classification is an automatic approach for classifying all pixels in the image but does not take into account the spatial information for the region of interest. On the other hand, region-growing methods take into account the neighbourhood pixels information. However, in region-growing methods, a pixel-group called 'points of interest' are needed to initialise the growing process. In this paper, we proposed an optimised pixel-based classification by the cooperation of region growing strategy. This original segmentation scheme is performed in two phases for the automatic recognition of white blood cells (WBC): the first is a learning step with colour characteristics of each pixel in the image. The second is a region growing application by classifying neighbouring pixels from pixels of interest extracted by the ultimate erosion technique. This process has proved that the cooperation allows obtaining a nucleus and cytoplasm segmentation as closer to what as expected in the reference images.

Improved surface extraction of multi-material components for single-source industrial X-ray computed tomography

The paper demonstrates a novel methodology for surface extraction of multi-material components (MMCs) on industrial X-ray computed tomography (CT) datasets. The methodology is based on a combination of fuzzy C-means clustering (FCM) and region growing (RG) methods. FCM, used as a pre-processing step, allows proper classification and improvement of different objects boundaries present on industrial X-ray CT datasets. Afterwards, application of RG method enables accurate segmentation of classified and improved X-ray CT datasets. The performance of presented approach has been tested on two CT datasets acquired on an industrial X-ray CT system NIKON XT H 225. It was also compared against two commercial industrial software VGStudio Max v3.1 and GOM Inspect v2018. Obtained results from application of the proposed approach show significant improvement in surface extraction of MMCs in CT datasets, especially in cases of low-density materials such as polymers. Verification has been conducted by obtaining reference measurements using contact coordinate measuring machine (CMM) Contura G2 by CARL ZEISS.

Workflow-centred open-source fully automated lung volumetry in chest CT

To develop a robust open-source method for fully automated extraction of total lung capacity (TLC) from computed tomography (CT) images and to demonstrate its integration into the clinical workflow. Using only open-source software, an algorithm was developed based on a region-growing method that does not require manual interaction. Lung volumes calculated from reconstructions with different kernels (TLCCT) were assessed. To validate the algorithm calculations, the results were correlated to TLC measured by pulmonary function testing (TLCPFT) in a subgroup of patients for which this information was available within 3 days of the CT examination. A total of 288 patients were analysed retrospectively. Manual review revealed poor segmentation results in 13 (4.5%) patients. In the validation subgroup, the correlation between TLCCT and TLCPFT was r=0.87 (p<0.001). Measurements showed excellent agreement between the two reconstruction kernels with an intraclass correlation coefficient (ICC) of 0.99. Calculation of the volumes took an average of 5 seconds (standard deviation: 3.72 seconds). Integration of the algorithm into the departments of the PACS environment was successful. A DICOM-encapsulated PDF document with measurements and an overlay of the segmentation results was sent to the PACS to allow the radiologists to detect false measurements. The algorithm developed allows fast and fully automated calculation of lung volume without any additional input from the radiologist. The algorithm delivers excellent segmentation in >95% of cases with significant positive correlations between lung volume on CT and TLC on PFT.

Automatic Tunnel Steel Arches Extraction Algorithm Based on 3D LiDAR Point Cloud.

Automation is an inevitable trend in the development of tunnel shotcrete machinery. Tunnel environmental perception based on 3D LiDAR point cloud has become a research hotspot. Current researches about the detection of tunnel point clouds focus on the completed tunnel with a smooth surface. However, few people have researched the automatic detection method for steel arches installed on a complex rock surface. This paper presents a novel algorithm to extract tunnel steel arches. Firstly, we propose a refined function for calibrating the tunnel axis by minimizing the density variance of the projected point cloud. Secondly, we segment the rock surface from the tunnel point cloud by using the region-growing method with the parameters obtained by analyzing the tunnel section sequence. Finally, a Directed Edge Growing (DEG) method is proposed to detect steel arches on the rock surface in the tunnel. Our experiment in the highway tunnels under construction in Changsha (China) shows that the proposed algorithm can effectively extract the points of the edge of steel arches from 3D LiDAR point cloud of the tunnel without manual assistance. The results demonstrated that the proposed algorithm achieved 92.1% of precision, 89.1% of recall, and 90.6% of the F-score.

Vascular segmentation of neuroimages based on a prior shape and local statistics

Fast and accurate extraction of vascular structures from medical images is fundamental for many clinical procedures. However, most of the vessel segmentation techniques ignore the existence of the isolated and redundant points in the segmentation results. In this study, we propose a vascular segmentation method based on a prior shape and local statistics. It could efficiently eliminate outliers and accurately segment thick and thin vessels. First, an improved vesselness filter is defined. This quantifies the likelihood of each voxel belonging to a bright tubular-shaped structure. A matching and connection process is then performed to obtain a blood-vessel mask. Finally, the region-growing method based on local statistics is implemented on the vessel mask to obtain the whole vascular tree without outliers. Experiments and comparisons with Frangi’s and Yang’s models on real magnetic-resonance-angiography images demonstrate that the proposed method can remove outliers while preserving the connectivity of vessel branches.

Measuring Brinell hardness indentation by using a convolutional neural network

For robust automatic measurement of Brinell hardness, we propose a novel measurement of the indentation diameter by using a convolutional neural network (CNN), which is a type of deep learning. In the Brinell hardness measurement, it is sometimes difficult to detect the indentation edges by image processing methods, due to the surface conditions or a change in the contrast with hardness levels. CNNs can automatically extract features required for object recognition of humans, hence we expect that the CNN can detect the indentation edges as robust as human operators, regardless of the surface conditions. We developed a CNN system to detect the indentation edges, and trained the CNN using the dataset combined with the indentation edge images and the position of edges given by a human operator. To verify the usefulness of the CNN method, we compared this method with measurement by a human operator and the region growing (RG) method, which is a simple region segmentation method in image processing. In the CNN method, indentation diameters and Brinell hardness were in good agreement with the manual measurement compared with the RG method, regardless of the hardness levels of test samples. Moreover, the CNN method enabled good measurement of a rough surface that is different from the dataset for training the CNN. Thus, our novel method is as robust as measurement by experienced operators and versatile in terms of independence of the hardness of samples or the characteristics of the surface conditions.

Fat-water separation based on Transition REgion Extraction (TREE).

To develop a method based on fat-water transition region extraction (TREE) for robust fat-water separation and quantification in challenging scenarios, including low signal-to-noise ratio (SNR), fast changing B0 field, and disjointed anatomies. In TREE method, the phasor solutions of each pixel were categorized into fat-dominant and water-dominant groups. The fat-water transition region was then extracted by detecting sudden changes in the phasor maps. The phasor solutions of the pixels in the transition region were solved by choosing the smoothest phasor combinations. For the remaining subregions, the phasor solution was then determined by all the surrounding transition region pixels. The proposed method was validated using various datasets, including some from the International Society for Magnetic Resonance in Medicine (ISMRM) 2012 Challenge. Quantitative score of proposed method (9936.8 of 10,000) is comparable to the winner (9951.9) of ISMRM 2012 Challenge. The total processing time was 179.3s for 15 datasets. Sagittal spine data with ~400mm field of view in head-foot direction were used to compare TREE with several representative region-growing methods. Results showed that the proposed method was robust under fast changing B0 field, disjointed anatomies and low SNR area. No apparent fat-water swap was observed in the low SNR (SNR ~ 10) dataset. Accurate proton density fat fraction results were also produced from the proposed method. A method based on fat-water transition region extraction was proposed for robust water-fat separation and fat fraction quantification. The method worked well in spatially disjointed objects, fast changing B0 field, and low SNR application.

Ground Moving Target Detection Using Beam-Doppler Image Feature Recognition

An innovative moving target detection approach termed as Beam-Doppler Image Feature Recognition (BDIFR) is introduced based on the distinction between moving target and ground clutter features in the beam-Doppler domain. A novel minimum-distance-based region-growing method is developed for radar target feature extraction and detection. The proposed BDIFR algorithm is advantageous over conventional space-time adaptive processing in detecting ground moving targets in inhomogeneous clutter environments since it does not require secondary training data for clutter estimation.

Single-Tree Detection in High-Resolution Remote-Sensing Images Based on a Cascade Neural Network

Traditional single-tree detection methods usually need to set different thresholds and parameters manually according to different forest conditions. As a solution to the complicated detection process for non-professionals, this paper presents a single-tree detection method for high-resolution remote-sensing images based on a cascade neural network. In this method, we firstly calibrated the tree and non-tree samples in high-resolution remote-sensing images to train a classifier with the backpropagation (BP) neural network. Then, we analyzed the differences in the first-order statistic features, such as energy, entropy, mean, skewness, and kurtosis of the tree and non-tree samples. Finally, we used these features to correct the BP neural network model and build a cascade neural network classifier to detect a single tree. To verify the validity and practicability of the proposed method, six forestlands including two areas of oil palm in Thailand, and four areas of small seedlings, red maples, or longan trees in China were selected as test areas. The results from different methods, such as the region-growing method, template-matching method, BP neural network, and proposed cascade-neural-network method were compared considering these test areas. The experimental results show that the single-tree detection method based on the cascade neural network exhibited the highest root mean square of the matching rate (RMS_Rmat = 90%) and matching score (RMS_M = 68) in all the considered test areas.

Altered White Matter Microstructure in the Corpus Callosum and Its Cerebral Interhemispheric Tracts in Adolescent Idiopathic Scoliosis: Diffusion Tensor Imaging Analysis.

Neural system was one of the important contributors to the etiopathogenesis of adolescent idiopathic scoliosis; additionally, the morphology of corpus callosum interconnecting both hemispheres of the brain was found to be altered morphologically. Our aim was to evaluate and compare the microstructural changes of the corpus callosum and its interhemispheric white matter fiber tracts interconnecting both cerebral hemispheres in patients with adolescent idiopathic scoliosis and matched controls using diffusion tensor imaging. Brain DTI was performed in 69 patients with adolescent idiopathic scoliosis (female, right thoracic/thoracolumbar curve) and 40 age-matched controls without adolescent idiopathic scoliosis (female). 2D and 3D segmentation of the corpus callosum were performed using a region-growing method, and the corpus callosum was further divided into 6 regions, including the rostrum, genu, anterior and posterior midbodies, isthmus, and splenium. The laterality index was calculated to quantify the asymmetry of the corpus callosum. Interhemispheric fiber tractography were performed using the Brodmann atlas. 2D ROI analysis revealed reduced fractional anisotropy in the genu and splenium (P = .075 and P = .024, respectively). Consistently reduced fractional anisotropy on the left sides of the genu and splenium was also found in 3D ROI analysis (P = .03 and P = .012, respectively). The laterality index analysis revealed a pseudo-right lateralization of the corpus callosum in adolescent idiopathic scoliosis. Interhemispheric fibers via the splenium interconnecting Brodmann 3, 1, and 2; Brodmann 17; and Brodmann 18 (corresponding to the primary somatosensory cortex and primary and secondary visual cortices) were also found to have reduced fractional anisotropy (P ≤ .05). Reduced fractional anisotropy was found in the genu and splenium of the corpus callosum and corresponding interhemispheric fiber tracts interconnecting the somatosensory and visual cortices via the splenium. Our results are suggestive of altered white matter microstructure within the brain of those with adolescent idiopathic scoliosis, which could be related to abnormal brain maturation during adolescence in adolescent idiopathic scoliosis and could possibly explain the previously documented somatosensory function impairment and visuo-oculomotor dysfunction in this condition.

Morphometric analysis of erythrocytes from patients with thalassemia using tomographic diffractive microscopy.

Complete blood count is the most common test to detect anemia, but it is unable to obtain the abnormal shape of erythrocytes, which highly correlates with the hematologic function. Tomographic diffractive microscopy (TDM) is an emerging technique capable of quantifying three-dimensional (3-D) refractive index (RI) distributions of erythrocytes without labeling. TDM was used to characterize optical and morphological properties of 172 erythrocytes from healthy volunteers and 419 erythrocytes from thalassemic patients. To efficiently extract and analyze the properties of erythrocytes, we developed an adaptive region-growing method for automatically delineating erythrocytes from 3-D RI maps. The thalassemic erythrocytes not only contained lower hemoglobin content but also showed doughnut shape and significantly lower volume, surface area, effective radius, and average thickness. A multi-indices prediction model achieved perfect accuracy of diagnosing thalassemia using four features, including the optical volume, surface-area-to-volume ratio, sphericity index, and surface area. The results demonstrate the ability of TDM to provide quantitative, hematologic measurements and to assess morphological features of erythrocytes to distinguish healthy and thalassemic erythrocytes.

APPLICATION OF EXTENDED KALMAN FILTER IN PERSISTANT SCATTERER INTERFEROMETRY TO ENHACE THE ACCURACY OF UNWRAPPING PROCESS

Abstract. In interferometry technique, phases have been modulated between 0-2π. Finding the number of integer phases missed when they were wrapped is the main goal of unwrapping algorithms. Although the density of points in conventional interferometry is high, this is not effective in some cases such as large temporal baselines or noisy interferograms. Due to existing noisy pixels, not only it does not improve results, but also it leads to some unwrapping errors during interferogram unwrapping. In PS technique, because of the sparse PS pixels, scientists are confronted with a problem to unwrap phases. Due to the irregular data separation, conventional methods are sterile. Unwrapping techniques are divided in to path-independent and path-dependent in the case of unwrapping paths. A region-growing method which is a path-dependent technique has been used to unwrap PS data. In this paper an idea of EKF has been generalized on PS data. This algorithm is applied to consider the nonlinearity of PS unwrapping problem as well as conventional unwrapping problem. A pulse-pair method enhanced with singular value decomposition (SVD) has been used to estimate spectral shift from interferometric power spectral density in 7*7 local windows. Furthermore, a hybrid cost-map is used to manage the unwrapping path. This algorithm has been implemented on simulated PS data. To form a sparse dataset, A few points from regular grid are randomly selected and the RMSE of results and true unambiguous phases in presented to validate presented approach. The results of this algorithm and true unwrapped phases were completely identical.

Usefulness of three-dimensional image navigation system for evaluation of hepatic artery before living donor liver transplantation: a case report

BackgroundThe evaluation of the hepatic vascular anatomy in living liver donors is increasingly being performed by three-dimensional (3D) computed tomography (CT) angiography. However, details of hepatic artery anatomy obtained by 3D CT angiography are not always superior to those obtained by angiography. Here, we report a case in which the 3D image navigation system helped to detect segment II, III, and IV arteries (A2, A3, and A4, respectively) that individually originated from the proper hepatic artery (PHA); this could not be detected by 3D CT angiography.Case presentationA 46-year-old man with end-stage primary biliary cirrhosis was admitted to our hospital for evaluation as a candidate for living donor liver transplantation. The patient’s younger sister, aged 43 years, was the only living donor candidate. The predicted left liver graft volume with the middle hepatic vein was found to be 403 mL using the region-growing method with 3D CT software. This volume was sufficiently large for the recipient because the standard liver volume of the recipient was 1095 mL. 3D CT angiography was performed twice but could not reveal the anatomical structure of the left and middle hepatic arteries. However, simulation using the region-growing method demonstrated individual branching off of A2, A3, and A4 from the PHA; conventional angiography demonstrated the same results. Each branch was approximately 1 mm in diameter, which was too small for secure anastomosis. Therefore, we selected the right liver graft for simplicity. The postoperative course of the donor and recipient was uneventful, and they were discharged on postoperative days 10 and 46, respectively.ConclusionsIn conclusion, reconstruction of the hepatic vasculature using the 3D software by region-growing method might be a useful adjunct for surgical planning in the evaluation of the hepatic arteries in living liver donors.

Relationship between Contrast Enhancement of the Perivascular Space in the Basal Ganglia and Endolymphatic Volume Ratio.

Purpose:We routinely obtain the endolymphatic hydrops (EH) image using heavily T2-weighted three dimensional-fluid attenuated inversion recovery (hT2w-3D-FLAIR) imaging at 4 hours after intravenous administration of a single-dose of gadolinium-based contrast media (IV-SD-GBCM). While repeating the examination, we speculated that the contrast enhancement of the perivascular space (PVS) in the basal ganglia might be related to the degree of EH. Therefore, the purpose of this study was to investigate the relationship between the endolymphatic volume ratio (%ELvolume) and the signal intensity of the PVS (SI-PVS).Materials and Methods:In 20 patients with a suspicion of EH, a heavily T2-weighted 3D-turbo spin echo sequence for MR cisternography (MRC) and an hT2w-3D-FLAIR as a positive perilymph image (PPI) were obtained at 4 hours after IV-SD-GBCM. The %ELvolume of the cochlea and the vestibule were measured on the previously reported HYDROPS2-Mi2 image. The PVS in the basal ganglia was segmented on MRC using a region-growing method. The PVS regions were copied and pasted onto the PPI, and the SI-PVS was measured. The larger value of the right and the left ears was employed as the %ELvolume, and the weighted average of both sides was employed as the SI-PVS. The correlation between the %ELvolume and the SI-PVS was evaluated.Result:There was a strong negative linear correlation between the %ELvolume of the cochlea and the SI-PVS (r = −0.743, P < 0.001); however, there was no significant correlation between the %ELvolume of the vestibule and the SI-PVS (r = −0.267, P = 0.256).Conclusion:There was a strong negative correlation between the cochlear %ELvolume and the SI-PVS. Contrast enhancement of PVS might be a biomarker of EH.

Use of Cardiac Computed Tomography for Ventricular Volumetry in Late Postoperative Patients with Tetralogy of Fallot.

BackgroundCardiac computed tomography (CT) has emerged as an alternative to magnetic resonance imaging (MRI) for ventricular volumetry. However, the clinical use of cardiac CT requires external validation.MethodsBoth cardiac CT and MRI were performed prior to pulmonary valve implantation (PVI) in 11 patients (median age, 19 years) who had undergone total correction of tetralogy of Fallot during infancy. The simplified contouring method (MRI) and semiautomatic 3-dimensional region-growing method (CT) were used to measure ventricular volumes.ResultsAll volumetric indices measured by CT and MRI generally correlated well with each other, except for the left ventricular end-systolic volume index (LV-ESVI), which showed the following correlations with the other indices: the right ventricular end-diastolic volume index (RV-EDVI) (r=0.88, p<0.001), the right ventricular end-systolic volume index (RV-ESVI) (r=0.84, p=0.001), the left ventricular end-diastolic volume index (LV-EDVI) (r=0.90, p=0.001), and the LV-ESVI (r=0.55, p=0.079). While the EDVIs measured by CT were significantly larger than those measured by MRI (median RV-EDVI: 197 mL/m2 vs. 175 mL/m2, p=0.008; median LV-EDVI: 94 mL/m2 vs. 92 mL/m2, p=0.026), no significant differences were found for the RV-ESVI or LV-ESVI.ConclusionThe EDVIs measured by cardiac CT were greater than those measured by MRI, whereas the ESVIs measured by CT and MRI were comparable. The volumetric characteristics of these 2 diagnostic modalities should be taken into account when indications for late PVI after tetralogy of Fallot repair are assessed.

An optimized pixel-based classification approach for automatic white blood cells segmentation

The pixel-based classification is an automatic approach for classifying all pixels in the image but does not take into account the spatial information for the region of interest. On the other hand, region-growing methods take into account the neighbourhood pixels information. However, in region-growing methods, a pixel-group called 'points of interest' are needed to initialise the growing process. In this paper, we proposed an optimised pixel-based classification by the cooperation of region growing strategy. This original segmentation scheme is performed in two phases for the automatic recognition of white blood cells (WBC): the first is a learning step with colour characteristics of each pixel in the image. The second is a region growing application by classifying neighbouring pixels from pixels of interest extracted by the ultimate erosion technique. This process has proved that the cooperation allows obtaining a nucleus and cytoplasm segmentation as closer to what as expected in the reference images.

Vision-based quantitative assessment of microcracks on reactor internal components of nuclear power plants

Ageing power facilities are increasingly susceptible to the onset of damage related to long exposure to stress, radiation, elevated temperatures and environmental conditions. One failure mechanism of particular concern is the onset of stress corrosion cracking. Currently, a technician manually measures the crack thicknesses at few points along a microcrack in a microscopic image, and the results are quantified by the Root Mean Square (RMS) of these measurements. This approach is time-consuming and subjective. In addition, the crack thickness is an important but difficult characteristic to accurately describe in a single numerical value, since it can vary considerably along the length of a crack. In this paper, a vision-based methodology is proposed for accurate quantification of microcracks that provides the thickness measurements for each pixel along the crack centreline. The proposed approach provides the minimum, maximum, mean, standard deviation, RMS and the histogram of crack thicknesses, which in most cases yields thousands of measurements. This approach provides more comprehensive insight regarding the condition of a microcrack. A region-growing method is used for segmenting microcracks from complex backgrounds. The microcrack thicknesses are then automatically computed along the lines orthogonal to the crack centreline. The fast marching method is used to accurately estimate the centreline of microcracks. Several real microcracks are examined to evaluate the performance of the proposed methodology.

Reproducibility and uptake time dependency of volume-based parameters on FDG-PET for lung cancer.

BackgroundVolume-based parameters, such as metabolic tumor volume (MTV) and total lesion glycolysis (TLG), on F-18 fluorodeoxyglucose (FDG) positron emission tomography (PET) are useful for predicting treatment response in nonsmall cell lung cancer (NSCLC). We aimed to examine intra- and inter-operator reproducibility to measure the MTV and TLG, and to estimate their dependency on the uptake time.MethodsFifty NSCLC patients underwent preoperative FDG-PET. After an injection of FDG, the whole body was scanned twice: at the early phase (61.4 ± 2.8 min) and delayed phase (117.7 ± 1.6 min). Two operators independently defined the tumor boundary using three different delineation methods: (1) the absolute SUV threshold method (MTVp and TLGp; p = 2.0, 2.5, 3.0, 3.5), (2) the fixed% SUVmax threshold method (MTVq% and TLGq%; q = 35, 40, 45), and (3) the adaptive region-growing method (MTVARG and TLGARG). Parameters were compared between operators and between phases.ResultsBoth the intra- and inter-operator reproducibility were high for all parameters using any method (intra-class correlation > 0.99 each). MTV3.0 and MTV3.5 resulted in a significant increase from the early to delayed phase (P < 0.05 for both), whereas MTV2.0 and MTV2.5 neither increased nor decreased (P = n.s.). All of the MTVq% values significantly decreased over time (P < 0.01), whereas MTVARG and TLG with any delineation method increased significantly (P < 0.05).ConclusionsHigh reproducibility of MTV and TLG was obtained by all of the methods used. MTV2.0 and MTV2.5 were the least sensitive to uptake time, and may be good alternatives when we compare images acquired with different uptake times, although applying constant uptake time is important for volume measurement.