A new level set model for unified medical image segmentation

In recent years, image segmentation based on level set formulations was encompassed by proficiently implementing various image processing applications, especially in case of reformative models for medical image segmentation. Existing level set methods under this criterion are not up to the mark in terms of noise, weak boundaries, and inhomogeneity. In this paper, we put forward a new formulation for selective segmentation with level set model using optimised fuzzy region clustering and implementation is carried out in two stages. In first, optimised fuzzy region clustering (OFR) is employed in which particle swarm optimisation technique is utilised at the initialisation of FCM clustering to obtain optimal cluster centres and then it is incorporated with region competition. While in second, implementing OFR clustering using level set methods. The proposed formulation is successfully implemented on some contemporary medicinal images. It is capable to identify the weak boundaries and segment the desired components of an image such as white substance, grey substance, cerebrospinal fluid, bone sarcoma, breast cancer and Alzheimer’s disease, etc. The visual results and quantitative analysis show its effectiveness, superiority and accuracy over the existing formulations.

Segmentation methods have been the subject of much research because they lead to many practical applications. Medical imaging is particularly concerned: it is indeed based on images of the inside of a patient's body (ultrasounds, x-rays, ...) on which it is desired to automatically identify the presence of abnormalities (for example the detection of micro-calcifications in a mammogram may reveal the presence of breast cancer. We can refer to for an overview of the possible applications of wavelets and segmentation problems in medical imaging); but image segmentation is also used in many other areas: shape tracking in video documents, detection of stellar objects in astronomical images, detection of fronts in satellite images for data assimilation in meteorology,....... In this paper, we study a New robust Level Set image segmentation method developed by Wavelet Transformation and Level set model. Proposed algorithm is verified for simulated images and applied for a medical image processing in the MATLAB environment.

Vascular segmentation plays an important role in medical image analysis. A novel technique for the automaticextraction of vascular trees from 2D medical images is presented, which combines Hessian-based multiscale filtering and a modified level set method. In the proposed algorithm, the morphological top-hat transformation is firstly adopted to attenuate background. Then Hessian-based multiscale filtering is used to enhance vascular structures by combining Hessian matrix with Gaussian convolution to tune the filtering response to the specific scales. Because Gaussian convolution tends to blur vessel boundaries, which makes scale selection inaccurate, an improved level set method is finally proposed to extract vascular structures by introducing an external constrained term related to the standard deviation of Gaussian function into the traditional level set. Our approach was tested on synthetic images with vascular-like structures and 2D slices extracted from real 3D abdomen magnetic resonance angiography (MRA) images along the coronal plane. The segmentation rates for synthetic images are above 95%. The results for MRA images demonstrate that the proposed method can extract most of the vascular structures successfully and accurately in visualization. Therefore, the proposed method is effective for the vascular tree extraction in medical images.

Multi-agent stochastic level set method in image segmentation

There is a challenge to segment the medical image which is often blurred and consists of noise. The objects to be segmented are always changing shape. Thus, there is a need to apply a method to automated segment well the objects for future analysis without any assumptions about the object's topology are made. In general, when performing pregnancy ultrasound scanning, obstetrician needs to find out the best position or angle of the foetus and freeze the scene. The obstetrician will click on the crown and the rump of the foetus to get the foetus length. The segmentation technique applied is level set method. A variational level set algorithm has been successfully implemented in medical image segmentation (Xray image, MRI image and ultrasound image). The results showed the level set contour evolved well on the low contrast and noise consisting medical image, especially the ultrasound image.

Level set methods have previously been successfully implemented in interface propagation for etching and deposition processes. In this article, the authors show that level set methods can be used to model solid phase epitaxial regrowth. The model incorporates orientation dependence of regrowth as found by Csepregi et al. [J. Appl. Phys. 49, 3906 (1978)]. The orientation dependent velocity data are taken from Csepregi’s paper and fitted to a polynomial function to give the growth velocity for level set methods. Simulations show the capability of our model in predicting the pinching of the corners in ⟨111⟩ direction and humplike shape in ⟨100⟩ direction. This is confirmed by the transmission electron microscope pictures from recent papers. This modeling holds special interest because of the different diffusivities of boron in amorphous and crystalline silicon (approximately five orders of magnitude difference) and because of the various defects forming at the pinching corners which could lead to higher leakage current in scaled devices. The level set model is implemented in FLOOPS.

Self-adaptive weighted level set evolution based on local intensity difference for parotid ducts segmentation

The level set method is a mathematical and computational, technique for tracking a moving interface over time. It can naturally handle topological changes such as merging or breaking interfaces. Intrinsic geometric properties of the interface, such as curvature and normal direction, are easily determined from the level set function p. There are many applications of the level set method, including kinetic crystal growth, epitaxial growth of thin films, image restoration, vortex dominated flows, and so forth. Most applications described in the growing literature on the applications of level sets advance the level set equation with explicit time integration. Hence, small CFL-respecting time steps are needed to maintain stability. In this thesis, an implicit level set method is introduced and applied to wildland firespread models, removing vulnerability to instability.

Regional level set method is a popular approach for image segmentation that uses inside and outside information of contour to extract object boundary. Unfortunately, in many cases, such method is not adequate to model complex textured objects. In this paper, we propose a natural texture image segmentation method which incorporates the pixel-level feature into region-level feature. The multiscale local structure operation is proposed as pixel-level feature to describe the texture structure of image. So the problems of multi-scale and rotation invariance of inhomogeneous texture are addressed by introducing multi-scale local structure operation into level set energy functional. Then, the global intensity information is extracted as the region-level feature and integrated with multi-scale local structure operation. Further, we propose a so-called vector level set method to obtain the segmentation results. Here, we extend the traditional regional level set model into the vector formulation so that the multi-scale local structure operation can be suitably combined with the global intensity information to achieve the more superior image segmentation performance than that of the traditional segmentation methods for texture images. Experiments on some synthesis texture images and real natural scene images demonstrate the excellent performance of the proposed method which successfully combines local structure information and global intensity information to extract the object boundary.

Medical image segmentation is a complex study due to its disadvantages such as noise, low-contrast, intensity inhomogeneity, and so on. A novel level set model was proposed in this study to segment medical images accurately. The kernel function used to determine the size of neighborhood of central pixel was modified by Laplace kernel function, which is insensitive to the choice of parameters and is more suitable for segmenting medical images. Compared with several state-of-the-art models, both visual and objective experiments can demonstrate the performance and superiority of the novel level set model.

Many medical and real images are suffered from intensity inhomogeneity and weak edges. For higher image segmentation quality, lots of level set-based methods have been proposed. Some of them however cannot take advantage of image gradient information. And severe intensity inhomogeneity and weak edges are not disposed properly. To address these problems, a new level set method integrated with local direction gradient information is presented in this paper. Firstly, according to the two assumptions on image intensity inhomogeneity adopted by many existing methods, a new pixel classification model based on image gradient is introduced. Secondly, we employ variational level set method combined with image spatial information, which improves the anti-noise capability of the proposed method. Finally, considering the gray gradients in homogeneous regions are close to constants, an improved diffusion process is incorporated into the level set function to make the evolving curves stay around true image edges. To verify our method, different testing images including synthetic images, magnetic resonance imaging (MRI) and real-world images are introduced. The image segmentation results demonstrate that our method can deal with the relatively severe intensity inhomogeneity and obtain the comparatively ideal segmentation results efficiently.

3D level set methods for evolving fronts on tetrahedral meshes with adaptive mesh refinement

An improved edge-based level set method combining local regional fitting information for noisy image segmentation

In recent years, generative adversarial networks (GANs) have gained tremendous popularity for various imaging related tasks such as artificial image generation to support AI training. GANs are especially useful for medical imaging-related tasks where training datasets are usually limited in size and heavily imbalanced against the diseased class. We present a systematic review, following the PRISMA guidelines, of recent GAN architectures used for medical image analysis to help the readers in making an informed decision before employing GANs in developing medical image classification and segmentation models. We have extracted 54 papers that highlight the capabilities and application of GANs in medical imaging from January 2015 to August 2020 and inclusion criteria for meta-analysis. Our results show four main architectures of GAN that are used for segmentation or classification in medical imaging. We provide a comprehensive overview of recent trends in the application of GANs in clinical diagnosis through medical image segmentation and classification and ultimately share experiences for task-based GAN implementations.

With increasing computer capabilities, numerical modeling of two-phase flows has developed significantly over the last few years. Although there are two main categories, namely 'one' fluid and 'two' fluid methods, the 'one' fluid methods are more commonly used for tracking or capturing the interface between two fluids. Level set (LS), volume-of-fluid (VOF), front tracking, marker-and-cell (MAC) and lattice Boltzmann (LB) methods are all 'one' fluid methods. It is clear that there is no perfect method; each method has advantages and disadvantages which make it more appropriate for one kind of problem than for others. For instance, a LS method will accurately compute the curvature and the normal to the interface, but tends to loss mass which is physically incorrect. On the other hand, a VOF method will conserve mass up to machine precision, but the computation of the curvature and normal to the interface is not as accurate. In order to minimize the disadvantages of these methods, several authors have used two or more methods together to model two-phase flows. This is the case for the CLSVOF (Couple Level Set Volume Of Fluid) method, where LS and VOF are coupled together in order to better capture the interface. In CLSVOF, the level set function is used to compute the interface curvature and normal to the interface, while the volume of fluid function is used to capture the interface. For two-phase flows in microchannels, surface tension forces play an important role in determining the dynamics of bubbles whereas gravitational forces are generally negligible. Also it is very important to consider the interaction between the boundaries and the fluids by prescribing or computing the correct contact angle between them. The commercial CFD code FLUENT allows the use of static constant angles, or the use of User Defined Functions (UDF) to compute the dynamic contact angles. It is inappropriate to use a static contact angle to model cases involving moving contact lines. For such cases a dynamic contact angle scheme should be implemented. In this study, FLUENT was used to model adiabatic and diabatic, time dependent two-phase flows. Since FLUENT already contains a VOF method, a LS method was implemented and coupled with VOF into FLUENT via UDFs. Furthermore, since the LS function, used to compute the surface tension force, ceases to be a signed distance to the interface even after one time step, a re-initialization equation was solved after each time step. This involved using a fifth order WENO (Weighted Essentially Non Oscillatory) scheme to discretize the space derivatives (otherwise oscillations of the interface occurred), and a first order Euler method for the time integration. In another part of the study, a 3D dynamic contact angle model based on volume fraction, interface reconstruction, and experimentally available advancing and receding static contact angles was also developed and implemented into FLUENT via UDFs. Several validations for the developed CLSVOF method and dynamic contact angle model are presented in this thesis, these includes a static bubble, a bubble rising in a stagnant liquid for Morton numbers ranging from 102 to 10-11, droplet deformation due to a vortex flow field, droplets spreading over a wall under the gravity effect and droplets sliding over a wall due to gravity. These validations demonstrated the high accuracy and the stability of our methods for modeling these phenomena. A heat and mass transfer model was also implemented into the commercial CFD code FLUENT for simulating of boiling (and condensation) heat transfer. Several simulations were presented with water and R134a as working fluids. The influence of the contact angle and the wall superheat was also studied.