Papers by Keyword: Image Segmentation

Abstract: In Laser Powder Bed Fusion (LPBF), ensuring reproducible part quality remains a ma-jor challenge despite the availability of high-resolution in-situ monitoring systems, such as ExposureOptical Tomography (EOT). While EOT provides detailed layer-wise optical information, most exist-ing approaches focus on single-layer analysis or real-time process control and do not exploit the fullvolumetric information contained in the acquired data.This work presents a modular framework for the volumetric reconstruction and post-process anal-ysis of EOT image data. Sequential EOT images are processed using volumetric component segmen-tation (VCS) and fused into a three-dimensional OT Image Cube, forming a central volumetric datastructure called LPBF Cube. Each voxel encodes spatial and radiometric information and can option-ally be augmented with additional process metadata.Based on this representation, high-resolution two-dimensional slices are rendered along arbitraryorientations using profile-based slicing strategies for planar, cylindrical, and complex geometries.These slices enable intuitive, part-level inspection of laser exposure history and spatial process vari-ations. The framework is validated using geometric and radiometric analyzes, demonstrating goodagreement with nominal CAD geometry and a clear correlation between EOT-derived emission val-ues, laser energy input, and local cross-sectional area.The proposed approach extends the use of EOT data beyond layer-wise monitoring toward com-prehensive, volumetric part inspection and provides a practical basis for geometry-aware quality as-sessment in LPBF, particularly for prototyping and post-build evaluation.

Abstract: A deep understanding on the intermixing of components in hybrid yarn or composite structures is decisive in order to develop hybrid structures with desired properties. This paper presents the development of a versatile procedure for the determination of the degree of fiber mixing in yarns and composites based on microscopy images auto-segmented by a neural network. The procedure is based on the quantification of blend irregularity values and blend homogeneity. For this purpose, functions of spatial point patterns analysis have been used to investigate the blend uniformity of yarn and composite cross sectional areas. The results show that the trained neural network model for segmentation of images has an accuracy of 92 %, indicating that the method is capable of accurately assessing the location of fibers in hybrid struc-tures. The results of the spatial point patterns analysis reveals a correlation between the blend value and the properties of yarns and composites. The proposed method provides a fast and reliable way to evaluate the hybrid structures, which could be used as a tool for quality control and process optimization.

Abstract: Brain tumors are developed as a result of unregulated and fast cell proliferation. It may result in death if not treated in the early stages. The imaging technology used to diagnose brain tumors is known as magnetic resonance imaging (MRI). Early detection of brain tumors is critical in medical practise in order to determine whether the tumor will progress to malignancy. For picture categorization, deep learning is a useful and effective method. Deep learning has been widely used in a variety of sectors, including medical imaging, because its application does not necessitate the expertise of a subject matter expert, but does necessitate a large amount of data and a variety of data in order to produce accurate classification results. The deep learning technique for image categorization is the convolutional neural network (CNN).In this research work , two different models are used to categorize brain tumors and their results were evaluated using performance metrics like accuracy and precision and the results were impressive

Abstract: White blood cell image segmentation provides the opportunity for medical experts to objectively diagnose the medical conditions of patients suffering from Leukemia, for example. Due to the rigorous nature of cell image acquisition (staining process and non-uniform illumination) efficient tools must be deployed to achieve the desired segmentation result. In this paper, a deep learning model is proposed together with a grab cut. The developed deep learning model provides an initial coarse segmentation of white blood cell images. However, the objective of this segmentation is to localize or identify regions of interest from white blood cell images. A bounding is generated from the localized cell image and then used to initiate an automatic cell image segmentation using grab cut. Results of the two publicly available datasets of white blood cell images are considered satisfactory on the proposed model.

Abstract: 3C silicon carbide is a semiconductor with remarkable properties, making it ideal for the development of long lasting devices, working in harsh environments and under high particle flows. The most significant obstacle to its wider diffusion is the presence of extended, bidimensional and linear defects in its crystal lattice. The purpose of this research is to automatically recognize defects from a TEM image by algorithm that calculates distances and angles.

Abstract: Due to their cost-efficiency and environmental friendliness, the demand of mechanical joining processes is constantly rising. However, the dimensioning and design of joints and suitable processes are mainly based on expert knowledge and few experimental data. Therefore, the performance of numerical and experimental studies enables the generation of optimized joining geometries. However, the manual evaluation of the results of such studies is often highly time-consuming. As a novel solution, image segmentation and machine learning algorithm provide methods to automate the analysis process. Motivated by this, the paper presents an approach for the automated analysis of geometrical characteristics using clinching as an example.

Abstract: Medical segmentation is the primary task performed to diagnosis the abnormalities in the human body. The brain is the complex organ and anatomical segmentation of brain tissues is a challenging task. In this paper, we used Enhanced Independent component analysis to perform the segmentation of gray matter. We used modified K means, Expected Maximization and Hidden Markov random field to provide better spatial correlation that overcomes in-homogeneity, noise and low contrast. Our objective is achieved in two steps (i) initially unwanted tissues are clipped from the MRI image using skull stripped Algorithm (ii) Enhanced Independent Component analysis is used to perform the segmentation of gray matter. We apply the proposed method on both T1w and T2w MRI to perform segmentation of gray matter at different noisy environments. We evaluate the the performance of our proposed system with Jaccard Index, Dice Coefficient and Accuracy. We further compared the proposed system performance with the existing frameworks. Our proposed method gives better segmentation of gray matter useful for diagnosis neurodegenerative disorders.

Abstract: MR Brain Image Segmentation is an important step in brain image analysis. It facilitates the automatic interpretation or diagnosis that helps in surgical planning, estimating the changes in the brain’s volume for various types of tissues, and recognizing different neural disorders. Many neurological disorders like epilepsy, Alzheimer’s, tumor, and cancer can be effectively quantified and analyzed by finding the volume of the brain tissues such as White Matter (WM), Gray Matter (GM), and Cerebro Spinal Fluids (CSF). In manual segmentation of brain MRIs physicians manually determines the boundaries of different objects of interest and it is time-consuming and difficult. Thus, several accurate automatic brain MRI segmentation techniques with different levels of complexity have been proposed. This paper proposes an advanced thresholding technique for the segmentation of brain MRIs based on the biologically inspired Ant Colony Optimization (ACO) algorithm. Here the texture features are assumed as heuristic data. The experimental results for the T1-weighted brain MRIs have shown high accuracy than the conventional such as Fuzzy C-Means (FCM), Expectation-Maximization (EM), Improved Bacterial Foraging Algorithm (IBFA), and Improved Particle Swarm Optimization (IPSO).

Abstract: With the rapid development of medical imaging technology, computer graphics and visualization technologies, virtual endoscopy technology emerged. It mainly includes 2D medical image segmentation, 3D image reconstruction, path planning and virtual roaming. However, the path planning of virtual endoscopy has become one of the obstacles in this field due to the high irregularity of the nasopharyngeal anatomy structure. In this study, the nasopharynx including meatus nasi, pharyngeal canal, maxillary sinus, frontal sinus, sphenoid sinus, and ethmoid sinus is segmented and 3D reconstructed using MR images. The key technology of virtual endoscopy - center path planning algorithm is implemented based on distance transform. Also, two improved algorithms of center path planning are proposed. One is the selection algorithm of branch path and the other is the extraction algorithm for complex path based on human-computer interaction. These two improved algorithms can not only allow the traditional path planning algorithm to handle multiple branching structure but make roaming path to start at any point. Our experimental results satisfied the needs of clinical practice.

Abstract: Recent research studies in the sphere of computer tomography are connected with the task of image analysis. Due to the fact that computed tomography images include artifacts, low contrast and different types of noises, researchers have to deal with a wide range of problems during the processing. There is a wide variety of preliminary processing methods which allow solving these problems. Obviously, the choice of these methods has a major impact on the result [1]. However, algorithm analysis of computed tomography images is not considered in the literature nowadays. This work presents an overview of the implementation approach of these methods.