Machine Learning Model for Multi-View Visualization of Medical Images

2020 ◽  
Author(s):  
Nitesh Pradhan ◽  
Vijaypal Singh Dhaka ◽  
Geeta Rani ◽  
Himanshu Chaudhary
Keyword(s):  
Magnetic Resonance Imaging ◽  
Machine Learning ◽  
Magnetic Resonance ◽  
Computerized Tomography ◽  
Imaging Techniques ◽  
Three Dimensional ◽  
Resonance Imaging ◽  
X Ray ◽  
Machine Learning Model ◽  
Tomography Scan

Abstract Imaging techniques such as X-ray, computerized tomography scan and magnetic resonance imaging are useful in the correct diagnosis of a disease or deformity in the organ. Two-dimensional imaging techniques such as X-ray give a clear picture of simple bone deformity but fail in visualizing multiple fractures in a bone. Moreover, these lack in providing a multi-angle view of a bone. Three-dimensional techniques such as computerized tomography scan and magnetic resonance imaging present a correct orientation of fracture geometry. Computerized tomography scan is a collection of multiple slices of an image. These slices provide a fair idea about a fracture but fail in the measurement of correct dimensions of a fractured fragment and to observe its geometry. It also exposes a patient with carcinogenic radiations. Magnetic resonance imaging induces a strong magnetic field. So, it becomes ineffective for organs containing metallic implants. The high cost of three-dimensional imaging techniques makes them inaccessible for economic weaker section of society. The limitations of two- and three-dimensional imaging techniques motivate researchers to propose an innovative machine learning model ‘CT slices to $3$-D convertor’ that accepts multiple slices of an image and yields a multi-dimensional view at all possible angles from 0 degree to 360 degree for an input image.

scholarly journals Early and late results of carotid endarterectomy: retrospective study of 70 operations

2001 ◽  
Vol 119 (6) ◽  
pp. 206-211 ◽  
Author(s):  
Eduardo Toledo de Aguiar ◽  
Alex Lederman ◽  
Celso Higutchi ◽  
Gerd Schreen
Keyword(s):  
Magnetic Resonance Imaging ◽  
Myocardial Infarction ◽  
Magnetic Resonance ◽  
Carotid Endarterectomy ◽  
Carotid Stenosis ◽  
Computerized Tomography ◽  
Late Results ◽  
Resonance Imaging ◽  
Operative Death ◽  
Tomography Scan

CONTEXT: Indications and results of carotid endarterectomy have been defined from clinical multicentric trials like the European Carotid Surgery Trialists, North-American Symptomatic Carotid Endarterectomy Trial and Asymptomatic Carotid Atherosclerosis Study. The patients included in these trials were highly selected, as were the surgeons performing the operations. Clinical practice is different but the same results should be achieved. OBJECTIVE: To study indications, technique, early and late results, and whether carotid endarterectomy has been performed in accordance with standards defined by multicentric trials. DESIGN: Retrospective case report study. SETTING: A tertiary care private hospital. PARTICIPANTS: 57 patients, on whom 70 carotid endarterectomies were performed over a 10-year period. The median age was 66.4 ± 7.8 years; 43 (75.4%) were male, 41 (71.9%) hypertensive, 36 (63.1%) current smokers and 24 (21.0%) had diabetes. Bilateral carotid stenosis was present in 31 (54.3%) patients, peripheral arterial occlusions in 32 (56.1%) and ischemic cardiopathy in 25 (43.1%). All patients had had angiography and 41 (71.9%) had also had a duplex-scan of neck arteries. Cerebral imaging via computerized tomography scan or magnetic resonance imaging was obtained for 36 patients. Patients were followed up over a period of one to 122 months. MAIN MEASUREMENTS: early and late post-operative death, early and late post-operative stroke, and recurrence of atheroma plaque and symptoms relative to carotid stenosis. RESULTS: There was one post-operative death (1.4%) caused by myocardial infarction and two early strokes (2.8%): a total complication rate of 4.2%. After 3 and 5 years, 95.4% and 81.3% of patients respectively were stroke-free and 72.8% and 67.3% were alive. There were four recurrences and two of them related to stroke. Forty-nine (70%) stenoses operated on were symptomatic. Brain infarction was detected in 59.2% of patients who underwent computerized tomography scan or magnetic resonance imaging. CONCLUSIONS: Carotid endarterectomy was done in accordance with international standards. The most frequent cause of late death was myocardial infarction, and recurrences were related to stroke. Patients should be followed up closely.

Current Status and New Developments in Breast Cancer Diagnosis and Detection

2013 ◽  
Vol 09 (01) ◽  
pp. 21 ◽  
Author(s):  
Demitrios Tzias ◽  
Elizabeth AM O’Flynn ◽  
Steven D Allen ◽  
A Robin M Wilson ◽  
◽  
...  
Keyword(s):  
Breast Cancer ◽  
Magnetic Resonance Imaging ◽  
Magnetic Resonance ◽  
Breast Imaging ◽  
Imaging Techniques ◽  
Current Status ◽  
Routine Practice ◽  
Resonance Imaging ◽  
High Frequency Ultrasound ◽  
X Ray

Despite extensive research into new ways of imaging the breast x-ray mammography and breast ultrasound, supplemented where necessary by magnetic resonance imaging, remain the techniques used for the vast majority of breast imaging for screening and the assessment of symptomatic breast problems. Recent advances in these technologies mean that these three techniques are highly effective for both detecting disease and for confirming normality. X-ray based imaging of the breast has been around now for 100 years but it is only in the last 10 years or so that digital technology developments have allowed for major advances in the efficacy of this technique. Digital breast tomosynthesis is currently the most promising technology as it has the potential to both improve detection of breast cancer and greatly reduce the numbers of false positive events. Technological advances in grey scale high frequency ultrasound imaging mean that it is now universally used in both symptomatic diagnosis and breast screening. Newer ultrasound techniques such as 3D imaging, Doppler analyses and elastography add some additional value but so far none of these has achieved their hoped for additional potential. Magnetic resonance imaging is currently the most sensitive imaging technique for the detection and characterisation of breast disease, but its cost remains a barrier to its more widespread use. Nuclear medicine techniques have a role is special circumstances but are yet to show that they should be used in routine practice. There are a large number of potential alternative new imaging techniques for the breast, but, as yet, none of these have shown any significant benefits over the current techniques. Dedicated breast computed tomography has perhaps the best promise but clinically effective breast imaging at present remains in the application and refinement of recent developments in digital mammography, ultrasound and magnetic resonance imaging.

scholarly journals Cartilage morphology at 3.0T: Assessment of three-dimensional magnetic resonance imaging techniques

2010 ◽  
Vol 32 (1) ◽  
pp. 173-183 ◽  
Author(s):  
Christina A. Chen ◽  
Richard Kijowski ◽  
Lauren M. Shapiro ◽  
Michael J. Tuite ◽  
Kirkland W. Davis ◽  
...  
Keyword(s):  
Magnetic Resonance Imaging ◽  
Magnetic Resonance ◽  
Imaging Techniques ◽  
Three Dimensional ◽  
Resonance Imaging ◽  
Magnetic Resonance Imaging Techniques

Brain Imaging in Functional Mental Disorders

1986 ◽  
Vol 31 (7) ◽  
pp. 675-680 ◽  
Author(s):  
Trevor Young ◽  
Peter Williamson
Keyword(s):  
Magnetic Resonance Imaging ◽  
Positron Emission Tomography ◽  
Magnetic Resonance ◽  
Brain Imaging ◽  
Computerized Tomography ◽  
Imaging Techniques ◽  
Emission Tomography ◽  
Resonance Imaging ◽  
Positron Emission ◽  
Eeg Topography

The application of brain imaging techniques to psychiatry is reviewed with respect to computerized tomography (CT), EEG topography, positron emission tomography (PET), and magnetic resonance imaging (MRI). While early computerized tomography studies have suggested structural abnormalities in schizophrenia, more recent studies have shown that most schizophrenics and patients with other disorders have normal CT scans. EEG topography and positron emission tomography have not been evaluated as fully as computerized tomography. However, preliminary studies indicate some functional abnormalities in schizophrenia and affective disorders compared to normal controls. Magnetic resonance imaging shows promise but has had only a limited application to date in psychiatry.

scholarly journals Predictability of myometrial, lower uterine segment and cervical invasion with 3D transvaginal ultrasonography and magnetic resonance imaging in endometrial cancer patients: a prospective cohort study

2018 ◽  
Vol 20 (3) ◽  
pp. 348 ◽  
Author(s):  
Nuri Yildirim ◽  
Bahadir Saatli ◽  
Semir Kose ◽  
Ceren Sancar ◽  
Cagnur Ulukus ◽  
...  
Keyword(s):  
Magnetic Resonance Imaging ◽  
Endometrial Cancer ◽  
Magnetic Resonance ◽  
Cancer Patients ◽  
Imaging Techniques ◽  
Three Dimensional ◽  
Lower Uterine Segment ◽  
Transvaginal Ultrasonography ◽  
Resonance Imaging ◽  
Sensitivity Specificity

Aims: The objective of this study is to identify the diagnostic performance of three-dimensional transvaginal ultrasonography (3D-US) and magnetic resonance imaging (MRI) in detecting myometrial, lower uterine segment and/or cervical invasion in endometrial cancer patients.Materials and methods: In this prospective study, 40 patients diagnosed with endometrial cancer were performed 3D-US and MRI, preoperatively. Deep myometrial, lower uterine segment and cervical invasion were evaluated subjectively and results were compared with the final histology as a gold standard.Results: Diagnostic accuracy of 3D-US for detecting deep myometrial, lower uterine segment and cervical invasion were 87.5%, 80% and 85%, respectively. The same results for MRI were 75%, 65% and 70%, respectively. For deep myometrial, lower uterine segment and/or cervical invasion in endometrial cancer, 3D-US had higher sensitivity, specificity, negative and positive predictive value and accuracy than MRI. The combination of these two imaging techniques had an increased sensitivity of detecting all parameters related with tumoral invasion but decreased specificity and the accuracy.Conclusion: 3D-US had better performance in detecting myometrial, lower uterine segment and/or cervical invasion than MRI in endometrial cancer patients. Combination of these techniques was not preferred according to this study.

scholarly journals A CNN-Based Autoencoder and Machine Learning Model for Identifying Betel-Quid Chewers Using Functional MRI Features

2021 ◽  
Vol 11 (6) ◽  
pp. 809
Author(s):  
Ming-Chou Ho ◽  
Hsin-An Shen ◽  
Yi-Peng Eve Chang ◽  
Jun-Cheng Weng
Keyword(s):  
Magnetic Resonance Imaging ◽  
Machine Learning ◽  
Magnetic Resonance ◽  
Functional Magnetic Resonance Imaging ◽  
Resting State ◽  
Learning Model ◽  
Betel Quid ◽  
Functional Magnetic Resonance ◽  
Resonance Imaging ◽  
Machine Learning Model

Betel quid (BQ) is one of the most commonly used psychoactive substances in some parts of Asia and the Pacific. Although some studies have shown brain function alterations in BQ chewers, it is virtually impossible for radiologists’ to visually distinguish MRI maps of BQ chewers from others. In this study, we aimed to construct autoencoder and machine-learning models to discover brain alterations in BQ chewers based on the features of resting-state functional magnetic resonance imaging. Resting-state functional magnetic resonance imaging (rs-fMRI) was obtained from 16 BQ chewers, 15 tobacco- and alcohol-user controls (TA), and 17 healthy controls (HC). We used an autoencoder and machine learning model to identify BQ chewers among the three groups. A convolutional neural network (CNN)-based autoencoder model and supervised machine learning algorithm logistic regression (LR) were used to discriminate BQ chewers from TA and HC. Classifying the brain MRIs of HC, TA controls, and BQ chewers by conducting leave-one-out-cross-validation (LOOCV) resulted in the highest accuracy of 83%, which was attained by LR with two rs-fMRI feature sets. In our research, we constructed an autoencoder and machine-learning model that was able to identify BQ chewers from among TA controls and HC, which were based on data from rs-fMRI, and this might provide a helpful approach for tracking BQ chewers in the future.

Sign in / Sign up

Export Citation Format

Share Document