• Journal of Trauma and Injury
• /
• v.34 no.2
• /
• pp.126-129
• /
• 2021

Fat embolism syndrome is a rare, but serious condition that occurs in patients with fractures of the long bones or who undergo orthopedic surgery. The main clinical features of fat embolism syndrome are an altered mental status, hypoxia, and petechial rash. Cerebral fat embolism is the most severe manifestation of fat embolism syndrome because it can lead to an altered mental status. The diagnosis of cerebral fat embolism is clinical, but brain magnetic resonance image (MRI) is helpful. There is usually an interval until symptoms, such as an altered mental status, develop after trauma. We report a case of cerebral fat embolism in which the patient's mental status deteriorated several hours after trauma and the initial findings were negative on diffusion-weighted MRI.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.12
• /
• pp.4816-4834
• /
• 2020

This paper proposes transfer learning and fine-tuning techniques for a deep learning model to detect three distinct brain tumors from Magnetic Resonance Imaging (MRI) scans. In this work, the recent YOLOv4 model trained using a collection of 3064 T1-weighted Contrast-Enhanced (CE)-MRI scans that were pre-processed and labeled for the task. This work trained with the partial 29-layer YOLOv4-Tiny and fine-tuned to work optimally and run efficiently in most platforms with reliable performance. With the help of transfer learning, the model had initial leverage to train faster with pre-trained weights from the COCO dataset, generating a robust set of features required for brain tumor detection. The results yielded the highest mean average precision of 93.14%, a 90.34% precision, 88.58% recall, and 89.45% F1-Score outperforming other previous versions of the YOLO detection models and other studies that used bounding box detections for the same task like Faster R-CNN. As concluded, the YOLOv4-Tiny can work efficiently to detect brain tumors automatically at a rapid phase with the help of proper fine-tuning and transfer learning. This work contributes mainly to assist medical experts in the diagnostic process of brain tumors.

• Journal of Multimedia Information System
• /
• v.9 no.1
• /
• pp.21-32
• /
• 2022

Early-stage diagnosis of Alzheimer's Disease (AD) from Cognitively Normal (CN) patients is crucial because treatment at an early stage of AD can prevent further progress in the AD's severity in the future. Recently, computer-aided diagnosis using magnetic resonance image (MRI) has shown better performance in the classification of AD. However, these methods use a traditional machine learning algorithm that requires supervision and uses a combination of many complicated processes. In recent research, the performance of deep neural networks has outperformed the traditional machine learning algorithms. The ability to learn from the data and extract features on its own makes the neural networks less prone to errors. In this paper, a dense neural network is designed for binary classification of Alzheimer's disease. To create a classifier with better results, we studied result of different activation functions in the prediction. We obtained results from 5-folds validations with combinations of different activation functions and compared with each other, and the one with the best validation score is used to classify the test data. In this experiment, features used to train the model are obtained from the ADNI database after processing them using FreeSurfer software. For 5-folds validation, two groups: AD and CN are classified. The proposed DNN obtained better accuracy than the traditional machine learning algorithms and the compared previous studies for AD vs. CN, AD vs. Mild Cognitive Impairment (MCI), and MCI vs. CN classifications, respectively. This neural network is robust and better.

• Journal of Multimedia Information System
• /
• v.9 no.3
• /
• pp.177-182
• /
• 2022

A deep neural network (DNN) includes variables whose values keep on changing with the training process until it reaches the final point of convergence. These variables are the co-efficient of a polynomial expression to relate to the feature extraction process. In general, DNNs work in multiple 'dimensions' depending upon the number of channels and batches accounted for training. However, after the execution of feature extraction and before entering the SoftMax or other classifier, there is a conversion of features from multiple N-dimensions to a single vector form, where 'N' represents the number of activation channels. This usually happens in a Fully connected layer (FCL) or a dense layer. This reduced 2D feature is the subject of study for our analysis. For this, we have used the FCL, so the trained weights of this FCL will be used for the weight-class correlation analysis. The popular DNN models selected for our study are ResNet-101, VGG-19, and GoogleNet. These models' weights are directly used for fine-tuning (with all trained weights initially transferred) and scratch trained (with no weights transferred). Then the comparison is done by plotting the graph of feature distribution and the final FCL weights.

• Investigative Magnetic Resonance Imaging
• /
• v.2 no.1
• /
• pp.113-119
• /
• 1998

Purpose : Because adenocarcinomas of the uterine cervix have lower 5-year survival rate than squamous cell carcinomas due to early lymph node metastasis and local extension, scrutiny of lymph node metastasis and local extension by radiologic examination is necessary in case of clinically diagnosed or suspected adenocarcinomas. The purpose of this study is to evaluate whether there are specific findings of these tumors, compared with squamous cell carcinomas, through the analysis of magnetic resonance (MR) imaging findings. Materials and Methods : Of 21 pathologically proven cervical adenocarcinomas, MR imaging findings of 18 tumors (histologic staging : two Ib, four IIa, two IIb, one IIIa, and one IIIb) were retrospectively analyzed and compared with those of 40 wquamous cell carcinoma in consecutive patients as a control group. T1-wetighted and fast spin echo T2-weighted images were obtained on the axial and sagittal planes, using a 1.5-T MR scanner. The largest diameter, location, signal intensity and degree of contrast enhancement contour, shape and longitudinal extent of the tumor and associated findings on MR image were analyzed. Results : The largest diameters of cervical adenocarcinomas ranged from 0.8 to 4.1 cm(mean, 2.2 cm). Of 18 adenocarcinomas, nine were of endocervical type. All adenocarcinomas were isointense to surrounding cervical stroma on T1-weighted images and hyperintense(homogeneous in ten, inhomogeneous in eight) on fast spin echo T2-weighted images. Adenocarcinomas enhanced on contrast study in all patients (homogeneous in six, inhomogeneous in 12 with hyperintese enhnacing rim in two). Eight adenocarcinomas had smooth contours and ten had irregular ones. The shape of adenocarcinoma was irregular in eight patients, barrel shape in six, papillary/polypod in three, and nodular in one. All adenocarcinomas involved lower half of the uterine cervix and six tumors extended up to the upper half. Pelvic lymph nodes of more than 1.5cm in diameter in two adenocarcinomas pateints and no detectable small pelvic lymph nodes on MR imaging in one patient were pathologically positive. Hydrometra was associated in two adenocarcinomas patients, and hematometra in one patient. Compared with squamous cell carcinomas, more frequent MR findings of endocervical type and barrel shape in cervical adenocarcinomas were statistically significant. Conclusion : Cervical adenocarcinomas had more frequent MR findings of endocervical type and barrel shape, compared with wquamous cell carcinomas. Adenocarcinoma of the uterine cervix may be suspected on MR imaging, when a cervical carcinoma is of barrel shape along the endocervical canal and tends to involve lymth nodes in earlier stages.

• Investigative Magnetic Resonance Imaging
• /
• v.2 no.1
• /
• pp.14-30
• /
• 1998

Manganese is an essential element in the body. It is mainly deposited in the liver and to a lesser degree in the basal ganglia of the brain and eliminated through the bile duct. Rapid turnover of managanese in the body makes it difficult to evaluate the manganese exposure in workers, esecially in those with irregular or intermittent exposure, like welders. Therefore, conventional biomarkers, including blood and urine manganese can provide only a limited information about the long-tern or cumulative exposure to manganese. Introduction of magnetic resonance imaging (MRI) made a progress in the assessment of manganese exposure in the medical conditions related to manganese accumulation, e. g. hepatic failure and long-term total parenteral nutrition. Manganese shortens spin-lattice(T1) relaxation time on MRI due to its paramagnetic property, resulting in high signal intensity (HSI) on T1-weighted image(T1W1) of MRI. Manganese deposition in the brain, therefore, can be visualizedas an HSI in the globus pallidus, the substantia nigra, the putamen and the pituitary. clinical and epidemiologic studies regarding the MRI findings in the cases of occupational and non-occupational manganese exposure were reviewed. relationships between HSI on T1W1 of MRI and age, gender, occupational manganese exposure, and neurological dysfunction were analysed. Relationships betwen biological exposure indices and HSI on MRE werealso reviewed. Literatures were reviewed to establish the relationships between HSI, Manganese deposition in the brain, pathologic findings, and neurological dysfunction. HSI on T1W1 of MRI reflects regional manganese deposition in the brain. This relationship enables an estimation of regional manganese deposition in the brain by analysing MR signal intensity. Manganese deposition in the brain can induce a neuronal loss in the basal ganglia but functional abnormality is supposed to be related to the cumulative exposure of manganese in the brain, use of brain MRI for the assessment of exposure in a group of workers seems to be hardly rationalized, while ti can be a useful adjunct for the evaluation of manganese exposure int he cases with suspected manganese-related health problems.

• Journal of Digital Convergence
• /
• v.12 no.2
• /
• pp.439-446
• /
• 2014

To evaluate the effects of an artifact by metal material for orthodontics in Magnetic Resonance Image (MRI) examination, wires and brackets used in orthodontics were selected and compared. Using a head coil, a $T_2$-weighted image, $T_1$-weighted image and FLAIR image were obtained. With obtained images, the sizes of the artifacts were measured and compared using Image J Program. In the research, the material with the biggest artifact in the wires and brackets for orthodontics was stainless steel wire. In the future, selecting and developing metal for correction should be considered also in other fields along with the purpose of orthodontics.

• Journal of Biomedical Engineering Research
• /
• v.30 no.6
• /
• pp.461-468
• /
• 2009

Magnetic resonance electrical impedance tomography (MREIT) enables us to perform high-resolution conductivity imaging of an electrically conducting object. Injecting low-frequency current through a pair of surface electrodes, we measure an induced magnetic flux density using an MRI scanner and this requires a sophisticated MR phase imaging method. Applying a conductivity image reconstruction algorithm to measured magnetic flux density data subject to multiple injection currents, we can produce multi-slice cross-sectional conductivity images. When there exists a local region of fat, the well-known chemical shift phenomenon produces misalignments of pixels in MR images. This may result in artifacts in magnetic flux density image and consequently in conductivity image. In this paper, we investigate chemical shift artifact correction in MREIT based on the well-known three-point Dixon technique. The major difference is in the fact that we must focus on the phase image in MREIT. Using three Dixon data sets, we explain how to calculate a magnetic flux density image without chemical shift artifact. We test the correction method through imaging experiments of a cheese phantom and postmortem canine head. Experimental results clearly show that the method effectively eliminates artifacts related with the chemical shift phenomenon in a reconstructed conductivity image.

• Investigative Magnetic Resonance Imaging
• /
• v.8 no.1
• /
• pp.42-49
• /
• 2004

Purpose : The quantization noise in magnetic resonance imaging (MRI) systems is analyzed. The signal-to-quantization noise ratio (SQNR) in the reconstructed image is derived from the level of quantization in the signal in spatial frequency domain. Based on the derived formula, the SQNRs in various main magnetic fields with different receiver systems are evaluated. From the evaluation, the quantization noise could be a major noise source determining overall system signal-to-noise ratio (SNR) in high field MRI system. A few methods to reduce the quantization noise are suggested. Materials and methods : In Fourier imaging methods, spin density distribution is encoded by phase and frequency encoding gradients in such a way that it becomes a distribution in the spatial frequency domain. Thus the quantization noise in the spatial frequency domain is expressed in terms of the SQNR in the reconstructed image. The validity of the derived formula is confirmed by experiments and computer simulation. Results : Using the derived formula, the SQNRs in various main magnetic fields with various receiver systems are evaluated. Since the quantization noise is proportional to the signal amplitude, yet it cannot be reduced by simple signal averaging, it could be a serious problem in high field imaging. In many receiver systems employing analog-to-digital converters (ADC) of 16 bits/sample, the quantization noise could be a major noise source limiting overall system SNR, especially in a high field imaging. Conclusion : The field strength of MRI system keeps going higher for functional imaging and spectroscopy. In high field MRI system, signal amplitude becomes larger with more susceptibility effect and wider spectral separation. Since the quantization noise is proportional to the signal amplitude, if the conversion bits of the ADCs in the receiver system are not large enough, the increase of signal amplitude may not be fully utilized for the SNR enhancement due to the increase of the quantization noise. Evaluation of the SQNR for various systems using the formula shows that the quantization noise could be a major noise source limiting overall system SNR, especially in three dimensional imaging in a high field imaging. Oversampling and off-center sampling would be an alternative solution to reduce the quantization noise without replacement of the receiver system.

• Investigative Magnetic Resonance Imaging
• /
• v.10 no.2
• /
• pp.108-116
• /
• 2006

Purpose : High-resolution spiral-scan imaging is performed at 3 Tesla MRI system. Since the gradient waveforms for the spiral-scan imaging have lower slopes than those for the Echo Planar Imaging (EPI), they can be implemented with the gradient systems having lower slew rates. The spiral-scan imaging also involves less eddy currents due to the smooth gradient waveforms. The spiral-scan imaging method does not suffer from high specific absorption rate (SAR), which is one of the main obstacles in high field imaging for rf echo-based fast imaging methods such as fast spin echo techniques. Thus, the spiral-scan imaging has a great potential for the high-speed imaging in high magnetic fields. In this paper, we presented various high-resolution images obtained by the spiral-scan methods at 3T MRI system for various applications. Materials and Methods : High-resolution spiral-scan imaging technique is implemented at 3T whole body MRI system. An efficient and fast higher-order shimming technique is developed to reduce the inhomogeneity, and the single-shot and interleaved spiral-scan imaging methods are developed. Spin-echo and gradient-echo based spiral-scan imaging methods are implemented, and image contrast and signal-tonoise ratio are controlled by the echo time, repetition time, and the rf flip angles. Results : Spiral-scan images having various resolutions are obtained at 3T MRI system. Since the absolute magnitude of the inhomogeneity is increasing in higher magnetic fields, higher order shimming to reduce the inhomogeneity becomes more important. A fast shimming technique in which axial, sagittal, and coronal sectional inhomogeneity maps are obtained in one scan is developed, and the shimming method based on the analysis of spherical harmonics of the inhomogeneity map is applied. For phantom and invivo head imaging, image matrix size of about $100{\times}100$ is obtained by a single-shot spiral-scan imaging, and a matrix size of $256{\times}256$ is obtained by the interleaved spiral-scan imaging with the number of interleaves of from 6 to 12. Conclusion : High field imaging becomes increasingly important due to the improved signal-to-noise ratio, larger spectral separation, and the higher BOLD-based contrast. The increasing SAR is, however, a limiting factor in high field imaging. Since the spiral-scan imaging has a very low SAR, and lower hardware requirements for the implementation of the technique compared to EPI, it is suitable for a rapid imaging in high fields. In this paper, the spiral-scan imaging with various resolutions from $100{\times}100$ to $256{\times}256$ by controlling the number of interleaves are developed for the high-speed imaging in high magnetic fields.