Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining Practical Machine Learning Tools and Techniques

eMedicine创建于1996年，由近万名临床医师作为作者或编辑参与此临床医学知识库的建设，其中编辑均是来自美国哈佛、耶鲁、斯坦福、芝加哥、德克萨斯、加州大学等各分校医学院的教授或副教授。

eMedicine

The core mission of National Institute of Neurological Disorders and Stroke (NINDS) is twofold. First, NINDS seeks fundamental knowledge about the brain and nervous system. Second, NINDS aims to use that knowledge to reduce the burden of neurological diseases. In support of its mission, NINDS performs and funds basic, translational, and clinical neuroscience research on more than 600 neurological diseases, including genetic diseases (e.g. Huntington’s disease; muscular dystrophy), developmental disorders (e.g. cerebral palsy), neurodegenerative diseases (e.g. Parkinson’s disease; Alzheimer’s disease; multiple sclerosis), metabolic diseases (e.g. Gaucher’s disease), cerebrovascular diseases (e.g. stroke; vascular dementia), trauma (e.g. spinal cord and head injury), convulsive disorders (e.g. epilepsy), infectious diseases (e.g. AIDS dementia) and brain tumors. Common marmosets (Callithrix jacchus) offer unique, powerful advantages to both components of the NINDS mission. In support of the first component, marmosets are particularly well suited for neuroanatomical and functional brain studies, as their brains retain the typical anatomical and functional organization of the primate brain. A major advantage is that the marmoset is a lissencephalic primate, which greatly facilitates the mapping of functional brain areas by neuroimaging techniques, such as fMRI and optical imaging, as well as by electrophysiology, with high spatial resolution. In support of the second component, marmosets are excellent models of neurological disorders. Unlike rodents, marmosets are outbred and every individual is genetically different. Further, the marmoset brain has a gray-to-white matter ratio comparable to humans, which strongly facilitates modeling diseases such as multiple sclerosis and small vessel disease. The species also exhibits the breadth of cognitive sophistication that distinguishes primates from other taxonomic groups. Finally, geneedited marmosets can be generated with an intergeneration time and establishment of transgenic lines 2-3 times faster than other primate species, which makes marmosets be the ideal primate species for the development of genetically engineered lines. For all of the above reasons, marmosets are poised to be a central player to advance the core mission of the NINDS.

National Institute of Neurological Disorders and Stroke

Deep neural networks for automatic detection of osteoporotic vertebral fractures on CT scans

Lumbar vertebral fractures vary greatly in types and causes and usually result from severe trauma or pathological
conditions such as osteoporosis. Lumbar wedge compression fractures are amongst the most common ones where
the vertebra is severely compressed forming a wedge shape and causing pain and pressure on the nerve roots
and the spine. Since vertebral segmentation is the first step in any automated diagnosis task, we present a fully
automated method for robustly localizing and segmenting the vertebrae for preparation of vertebral fracture
diagnosis. Our segmentation method consists of five main steps towards the CAD(Computer-Aided Diagnosis)
system: 1) Localization of the intervertebral discs. 2) Localization of the vertebral skeleton. 3) Segmentation
of the individual vertebra. 4) Detection of the vertebrae center line and 5) Detection of the vertebrae major
boundary points. Our segmentation results are promising with an average error of 1.5mm (modified Hausdorff
distance metric) on 50 clinical CT cases i.e. a total of 250 lumbar vertebrae. We also present promising
preliminary results for automatic wedge compression fracture diagnosis on 15 cases, 7 of which have one or more
vertebral compression fracture, and obtain an accuracy of 97.33%.

/pdf/automatic-lumbar-vertebra-segmentation-from-clinical-ct-for-1w2vd9g693.pdf

Automatic lumbar vertebra segmentation from clinical CT for wedge compression fracture diagnosis

A Computer-aided diagnosis (CAD) system aims to facilitate characterization and quantification of abnormalities as well as minimize interpretation errors caused by tedious tasks of image screening and radiologic diagnosis. The system usually consists of segmentation, feature extraction and diagnosis, and segmentation significantly affects the diagnostic performance. In this paper, we propose an automatic segmentation method that extracts the spinal cord and the dural sac from T2-weighted sagittal magnetic resonance (MR) images of lumbar spine without the need of any human intervention. Our method utilizes a gradient vector flow (GVF) field to find the candidate blobs and performs a connected component analysis for the final segmentation. MR Images from fifty two subjects were employed for our experiments and the segmentation results were quantitatively compared against reference segmentation by two medical specialists in terms of a mutual overlap metric. The experimental results showed that, on average, our method achieved a similarity index of 0.7 with a standard deviation of 0.0571 that indicated a substantial agreement. We plan to apply this segmentation method to computer-aided diagnosis of many lumbar-related pathologies.

/pdf/automatic-segmentation-of-the-spinal-cord-and-the-dural-sac-2uk16528c0.pdf

Automatic segmentation of the spinal cord and the dural sac in lumbar MR images using gradient vector flow field

Purpose Detection of abnormal discs from clinical T2-weighted MR Images. This aids the radiologist as well as subsequent CAD methods in focusing only on abnormal discs for further diagnosis. Furthermore, it gives a degree of confidence about the abnormality of the intervertebral discs that helps the radiologist in making his decision. Materials and methods We propose a probabilistic classifier for the detection of abnormality of intervertebral discs. We usethreefeaturestolabelabnormaldiscsthatincludeappearance, location, and context. We model the abnormal disc appearance with a Gaussian model, the location with a 2D Gaussian model, and the context with a Gaussian model for the distance between abnormal discs. We infer on the middle slice of the T2-weighted MRI volume for each case. These MRI scans are specific for the lumbar area. We obtain our gold standard for the ground truth from our collaborating radiologist group by having the clinical diagnosis report for each case. Results We achieve over 91% abnormality detection accuracy in a cross-validation experiment with 80 clinical cases. The experiment runs ten rounds; in each round, we randomly leave 30 cases out for testing and we use the other 50 cases for training. Conclusion Weachievehighaccuracyfordetectionofabnormal discs using our proposed model that incorporates disc appearance,location,andcontext.Weshowtheextendability

/pdf/computer-aided-diagnosis-of-lumbar-disc-pathology-from-dvi32l7zsk.pdf

Computer-aided diagnosis of lumbar disc pathology from clinical lower spine MRI

The spinal cord is a vital organ that serves as the only communication link between the brain and the various parts of the body. It is vulnerable to traumatic spinal cord injury and various diseases such as tumors, infections, inflammatory diseases and degenerative diseases. The exact segmentation and localization of the spinal cord are essential to effective clinical management of such conditions. In recent years, due to the advances in imaging technology, the structure of internal organs and tissues can be captured accurately, and various abnormalities are diagnosed based on scanned images. In this paper, we present an unsupervised segmentation method that automatically extracts the spinal canal in the sagittal plane of magnetic resonance (MR) images. This segmentation method based on a novel saliency-driven attention model and a standard active contour model requires no human intervention and no training. Experiments based on 60 patients' data show that this procedure performs segmentation robustly, achieving the Dice's similarity index of 0.71 between the segmentation by our model and reference segmentation, as compared to the Dice's similarity index of 0.90 between two observers.

An automatic segmentation method of the spinal canal from clinical MR images based on an attention model and an active contour model

In this paper we propose a robust and fully automated lumbar herniation diagnosis system based on clinical MRI data which will not only aid a radiologist to make a decision with increased confidence, but will also reduce the time needed to analyze each case. Our method is based on three steps: 1) We automatically label the five lumbar intervertebral discs in a sagittal MRI slice using a probabilistic model and then extract an ROI for each disc using an Active Shape Model. 2) We generate relevant intensity and texture features from each disc ROI. 3) We construct five different classifiers (SVM, PCA+LDA, PCA+Naive Bayes, PCA+QDA, PCA+SVM) and combine them in a majority voting scheme. We perform 5-fold cross-validation experiments and achieve an accuracy of 94.85%, specificity of 95.9% and sensitivity of 92.45% for 35 clinical cases, i.e. a total of 175 lumbar intervertebral discs.

Gurmeet Dhillon

Papers

Automatic lumbar vertebra segmentation from clinical CT for wedge compression fracture diagnosis

Automatic segmentation of the spinal cord and the dural sac in lumbar MR images using gradient vector flow field

Computer-aided diagnosis of lumbar disc pathology from clinical lower spine MRI

An automatic segmentation method of the spinal canal from clinical MR images based on an attention model and an active contour model

Computer-aided diagnosis for lumbar mri using heterogeneous classifiers