Review on 2D and 3D MRI Image Segmentation Techniques

Open-source magnetic resonance imaging: Improving access, science, and education through global collaboration

Open-source practices and resources in magnetic resonance imaging (MRI) have increased substantially in recent years. This trend started with software and data being published open-source and, more recently, open-source hardware designs have become increasingly available. These developments towards a culture of sharing and establishing nonexclusive global collaborations have already improved the reproducibility and reusability of code and designs, while providing a more inclusive approach, especially for low-income settings. Community-driven standardization and documentation efforts are further strengthening and expanding these milestones. The future of open-source MRI is bright and we have just started to discover its full collaborative potential. In this review we will give an overview of open-source software and open-source hardware projects in human MRI research.

New Insights into Boxer's Knuckle Injury of the Little Finger

BACKGROUND

The original description of boxer's knuckle injury of the fifth ray mentions that the injury occurs between the extensor digitorum communis (EDC) and the extensor digiti minimi (EDM). Subsequent reports claim similar findings. Anatomical studies show that the EDC of the fifth ray is absent in most patients, while the EDM is generally composed of two slips. We present a modification of the current description of boxer's knuckle injury of the little finger based on the correlation between advanced preoperative 3D imaging and intraoperative findings.

METHODS

Five patients were investigated preoperatively using high-resolution ultrasound and 3D tendon reconstruction-based MR arthrography. Surgical exploration identified the lesion site relative to the EDM and EDC.

RESULTS

All patients had two slips of the EDM and no EDC to the fifth ray. The injury appeared as a longitudinal tear of the EDM between its two slips. The mean gap was 7.8 mm (range 4.5-10 mm) on the pathological side vs. 1.3 mm (range 1-2 mm) on the healthy contralateral side.

CONCLUSIONS

We believe that previous descriptions of boxer's knuckle of the fifth ray are inaccurate. High-resolution ultrasound and 3D reconstructions based on MR arthrography are reliable diagnostic tools allowing to locate the injury with precision.

An efficient 3D reconstruction method based on WT-TV denoising for low-dose CT images

BACKGROUND

In order to reduce the impact of CT radiation, low-dose CT is often used, but low-dose CT will bring more noise, affecting image quality and subsequent 3D reconstruction results.

OBJECTIVE

The study presents a reconstruction method based on wavelet transform-total variation (WT-TV) for low-dose CT.

METHODS

First, the low-dose CT images were denoised using WT and TV denoising methods. The WT method could preserve the features, and the TV method could preserve the edges and structures. Second, the two sets of denoised images were fused so that the features, edges, and structures could be preserved at the same time. Finally, FBP reconstruction was performed to obtain the final 3D reconstruction result.

RESULTS

The results show that The WT-TV method can effectively denoise low-dose CT and improve the clarity and accuracy of 3D reconstruction models.

CONCLUSION

Compared with other reconstruction methods, the proposed reconstruction method successfully addressed the issue of low-dose CT noising by further denoising the CT images before reconstruction. The denoising effect of low-dose CT images and the 3D reconstruction model were compared via experiments.

Coarse X-ray Lumbar Vertebrae Pose Localization and Registration Using Triangulation Correspondence

Plain film X-ray scanners are indispensable for medical diagnostics and clinical procedures. This type of device typically produces two radiographic images of the human spine, including the anteroposterior and lateral views. However, these two photographs presented perspectives that were distinct. The proposed procedure consists of three fundamental steps. For automated cropping, the grayscale lumbar input image was initially projected vertically using its vertical pattern. Then, Delaunay triangulation was performed with the SURF features serving as the triangle nodes. The posture area of the vertebrae was calculated by utilizing the edge density of each node. The proposed method provided an automated estimation of the position of the human lumbar vertebrae, thereby decreasing the radiologist’s workload, computing time, and complexity in a variety of bone-clinical applications. Numerous applications can be supported by the results of the proposed method, including the segmentation of lumbar vertebrae pose, bone mineral density examination, and vertebral pose deformation. The proposed method can estimate the vertebral position with an accuracy of 80.32 percent, a recall rate of 85.37 percent, a precision rate of 82.36%, and a false-negative rate of 15.42 percent.

AI-Assisted Diagnosis and Decision-Making Method in Developing Countries for Osteosarcoma

Osteosarcoma is a malignant tumor derived from primitive osteogenic mesenchymal cells, which is extremely harmful to the human body and has a high mortality rate. Early diagnosis and treatment of this disease is necessary to improve the survival rate of patients, and MRI is an effective tool for detecting osteosarcoma. However, due to the complex structure and variable location of osteosarcoma, cancer cells are highly heterogeneous and prone to aggregation and overlap, making it easy for doctors to inaccurately predict the area of the lesion. In addition, in developing countries lacking professional medical systems, doctors need to examine mass of osteosarcoma MRI images of patients, which is time-consuming and inefficient, and may result in misjudgment and omission. For the sake of reducing labor cost and improve detection efficiency, this paper proposes an Attention Condenser-based MRI image segmentation system for osteosarcoma (OMSAS), which can help physicians quickly locate the lesion area and achieve accurate segmentation of the osteosarcoma tumor region. Using the idea of AttendSeg, we constructed an Attention Condenser-based residual structure network (ACRNet), which greatly reduces the complexity of the structure and enables smaller hardware requirements while ensuring the accuracy of image segmentation. The model was tested on more than 4000 samples from two hospitals in China. The experimental results demonstrate that our model has higher efficiency, higher accuracy and lighter structure for osteosarcoma MRI image segmentation compared to other existing models.

Towards Drug Delivery Control Using Iron Oxide Nanoparticles in Three-Dimensional Magnetic Resonance Imaging

The purpose of this paper was to detect and separate the cluster intensity provided by Iron oxide nanoparticles (IO-NPs), in the MRI images, to investigate the drug delivery effectiveness. IO-NPs were attached to the macrophages and inserted into the eye of the inflamed mouse’s calf. The low resolution of MRI and the tiny dimension of the IO-NPs made the situation challenging. IO-NPs serve as a marker, due to their strong intensity in the MRI, enabling us to follow the track of the macrophages. An image processing procedure was developed to estimate the position and the amount of IO-NPs spreading inside the inflamed mouse leg. A fuzzy Clustering algorithm was adopted to select the region of interest (ROI). A 3D model of the femoral region was used for the detection and then the extraction IO-NPs in the MRI images. The results achieved prove the effectiveness of the proposed method to improve the control process of targeted drug delivered. It helps in optimizing the treatment and opens a promising novel research axis for nanomedicine applications.

A Complex Chained P System Based on Evolutionary Mechanism for Image Segmentation

A new clustering membrane system using a complex chained P system (CCP) based on evolutionary mechanism is designed, developed, implemented, and tested. The purpose of CCP is to solve clustering problems. In CCP, two kinds of evolution rules in different chained membranes are used to enhance the global search ability. The first kind of evolution rules using traditional and modified particle swarm optimization (PSO) clustering techniques are used to evolve the objects. Another based on differential evolution (DE) is introduced to further improve the global search ability. The communication rules are adopted to accelerate the convergence and avoid prematurity. Under the control of evolution-communication mechanism, the CCP can effectively search for the optimal partitioning and improve the clustering performance with the help of the distributed parallel computing model. This proposed CCP is compared with four existing PSO clustering approaches on eight real-life datasets to verify the validity. The computational results on tested images also clearly show the effectiveness of CCP in solving image segmentation problems.

Age-dependent decrease in dental pulp cavity volume as a feature for age assessment: a comparative in vitro study using 9.4-T UTE-MRI and CBCT 3D imaging

Evaluation of secondary dentin formation is generally suitable for age assessment. We investigated the potential of modern magnetic resonance imaging (MRI) technology to visualize the dental pulp in direct comparison with cone beam computed tomography (CBCT). To this end, we examined 32 extracted human teeth (teeth 11–48 [FDI]) using 9.4-T ultrashort echo time (UTE)-MRI and CBCT (methods). 3D reconstruction was performed via both manual and semi-automatic segmentation (settings) for both methods in two runs by one examiner. Nine teeth were also examined by a second examiner. We evaluated the agreement between examiners, scan methods, and settings. CBCT was able to determine the pulp volume for all teeth. This was not possible for two teeth on MRI due to MRI artifacts. The mean pulp volume estimated by CBCT was consistently higher (~ 43%) with greater variability. With lower variability in its measurements, evaluation of pulp volume using the MRI method exhibited greater sensitivity to differences between settings (p = 0.016) and between examiners (p = 0.009). The interactions of single-rooted teeth and multi-rooted teeth and method or setting were not found to be significant. For examiner agreement, the mean pulp volumes were similar with overlapping measurements (ICC > 0.995). Suitable for use in age assessment is 9.4-T UTE-MRI with good reliability and lower variation than CBCT. For MRI, manual segmentation is necessary due to a more detailed representation of the interior of the pulp cavity. Since determination of pulp volume is expected to be systematically larger using CBCT, method-specific reference values are indispensable for practical age assessment procedures. The results should be verified under in vivo conditions in the future.

3D-MRI Brain Tumor Detection Model Using Modified Version of Level Set Segmentation Based on Dragonfly Algorithm

Accurate brain tumor segmentation from 3D Magnetic Resonance Imaging (3D-MRI) is an important method for obtaining information required for diagnosis and disease therapy planning. Variation in the brain tumor’s size, structure, and form is one of the main challenges in tumor segmentation, and selecting the initial contour plays a significant role in reducing the segmentation error and the number of iterations in the level set method. To overcome this issue, this paper suggests a two-step dragonfly algorithm (DA) clustering technique to extract initial contour points accurately. The brain is extracted from the head in the preprocessing step, then tumor edges are extracted using the two-step DA, and these extracted edges are used as an initial contour for the MRI sequence. Lastly, the tumor region is extracted from all volume slices using a level set segmentation method. The results of applying the proposed technique on 3D-MRI images from the multimodal brain tumor segmentation challenge (BRATS) 2017 dataset show that the proposed method for brain tumor segmentation is comparable to the state-of-the-art methods.

Semantic segmentation of the multiform proximal femur and femoral head bones with the deep convolutional neural networks in low quality MRI sections acquired in different MRI protocols

Medical image segmentation is one of the most crucial issues in medical image processing and analysis. In general, segmentation of the various structures in medical images is performed for the further image analyzes such as quantification, assessment, diagnosis, prognosis and classification. In this paper, a research study for the 2D semantic segmentation of the multiform, both spheric and aspheric, femoral head and proximal femur bones in magnetic resonance imaging (MRI) sections of the patients with Legg-Calve-Perthes disease (LCPD) with the deep convolutional neural networks (CNNs) is presented. In the scope of the proposed study, bilateral hip MRI sections acquired in coronal plane were used. The main characteristic of the MRI sections that were used is to be low quality images which were obtained in different MRI protocols by using 3 different MRI scanners with 1.5 T imaging capability. In performance evaluations, promising segmentation results were achieved with deep CNNs in low quality MRI sections acquired in different MRI protocols. A success rate about 90% was observed in semantic segmentation of the multiform femoral head and proximal femur bones in a total of 194 MRI sections obtained from 33 MRI sequences of 13 patients with deep CNNs.

Autofocusing and image fusion for multi-focus plankton imaging by digital holographic microscopy

Digital holographic microscopy is becoming increasingly useful for the analysis of marine plankton. In this study, we investigate autofocusing and image fusion in digital holographic microscopy. We propose an area metric autofocusing method and an improved wavelet-based image fusion method. In the area metric autofocusing method, a hologram image is initially segmented into several plankton regions for focus plane detection, and an area metric is then applied to these regions. In the improved wavelet-based image fusion method, a marked map is introduced for labeling each plankton region with the order of refocus plane images that accounts for the most pixels. The results indicate that the area metric autofocusing method applied to each plankton region provides a higher depth resolution accuracy than a number of general autofocusing methods, and the mean accuracy increases by approximately 33%. The improved wavelet-based image fusion method can fuse more than nine reconstructed plane images at a time and effectively eliminate fringes and speckle noise, and the fused image is much clearer than that of a general wavelet-based method, a sparse decomposition method, and a pulse-coupled neural networks method. This work has practical value for plankton imaging using digital holographic microscopy.