1,721,144 research outputs found
A DETERMINISTIC AND DUCTILE SEGMENTATION ALGORITHM FOR MORPHOLOGIC MRI AND CTA IMAGES AND QUANTITATIVE ANALYSIS OF DYNAMIC SUSCEPTIBILITY-CONTRAST MAGNETIC RESONANCE IMAGING DATA
Full text linkThe work that is described in this thesis has been performed in collaboration
with the Research Unit in Brain Imaging and Neuropsychology (RUBIN) of the
Inter University Centre for Behavioural Neurosciences of Udine and Verona
(ICBN). The research group studies morphological and functional alterations of
the Brain in patients affected by psychiatric diseases as schizophrenia.
Several studies have been realized to investigate morphologic and functional
alterations in the Brain of patients affected by schizophrenia. Those studies are
usually performed on scans of the head acquired using structural Magnetic
Resonance Imaging (MRI) or functional MRI.
The morphological and functional differences described in scientific literature
between patients affected by schizophrenia and healthy controls are usually
very small and characterized by high variability. This fact depends on the high
inter-subject variability of the human brain, on the use of drugs by the patients
(that can affect the parameters), and on the diagnosis criteria of schizophrenia,
that include many different symptoms.
The implementation of procedures able to identify and analyze with high
accuracy and sensitivity the small differences that exist between patients
affected by schizophrenia and the healthy controls is a challenging problem at
the state of the art.
The first aim of the work described in this thesis is the ideation, the
implementation and the characterization of a fully automatic, robust, accurate
and ductile algorithm for the segmentation of the Brain, Cerebro Spinal Fluid,
Grey Matter and White Matter in T1 MRI of the head. In medical imaging,
segmentation can be defined as the identification of the boundaries of different
anatomical structures in the images. Segmentation algorithms are a key
component in medical imaging since they play a vital role in numerous
biomedical imaging applications. Depending on the state of the art that the
processes have reached, various methods have been realized to segment
specific anatomical structures.
The procedures to segment the Brain, Cerebro Spinal Fluid, Grey Matter and
White Matter in order to analyze the morphological alterations that are related
to psychiatric disease need to be very sensitive and accurate. Moreover, the
procedures need to be easy modified according to specific needs of the
research.
The Research Unit in Brain Imaging and Neuropsychology decided to
undertake the way of realizing a segmentation algorithm, rather than using one
of the available software, because those are considered unsatisfactory for the
research field. Moreover, the RUBIN desired to have an instrument of its own,
well known and easy to be modified according to specific needs. Ductility is
therefore a fundamental characteristic for the algorithm.
The realized algorithm is named Orao, it is fully automatic and is based on
iterative analyses of global and local intensity distributions, the application of
morphologic operators and the analysis of connectivity properties. It shows
excellent results in quantitative validation and comparison with the procedures
that are most used at the state of the art.
The second aim of the work is to test the ductility of the segmentation
algorithm through its application to various anatomical structures and medical
imaging acquisition techniques. The studied anatomical structures include the
Skull, Heart, Kidneys, Urinary Bladder, Urinary Tracts, Bone and a tumour of
the Brain. The studied medical imaging acquisition techniques include T1 and
T2 weighted MRI of the head and Computed Tomography Angiography of the
chest. Those applications are automatic or semiautomatic, depending on the
specific case, but it shall be noticed that the semi-automatic applications are
ideated in order to be easily automated.
The third aim of this work is the ideation, the implementation and the
characterization of a procedure to perform voxel by voxel analysis of Dynamic
Susceptibility Contrast MRI (DSC MRI). DSC MRI is a technique to perform
perfusion magnetic resonance using an exogenous tracer, such as
gadolinium, and is one of the most interesting techniques for the quantitative
study of the brain hemodynamics. The DSC MRI allows to quantify
important hemodynamic parameters that play an important role in the study of
several pathologies, such as cerebral tumours, ischemia or infarction,
epilepsy, but preliminary works suggest that this technique may provide
important clinical information on neuropsychiatric disorders, especially
dementia and schizophrenia.
Procedures that can compare voxel by voxel the brains of patients with the ones
of the healthy controls are still needed in Dynamic Susceptibility Contrast
MRI. A technique to perform DSC MRI analysis voxel by voxel could lead to
the identification of the anatomic regions majorly involved in various
pathologies as schizophrenia.
The fourth aim of the work is the analysis of local and global, morphological
and functional, alterations of the Brain, Grey Matter, White Matter and Cerebro
Spinal Fluid (CSF) in patients affected by schizophrenia using the procedures
realized. First morphological alterations are studied through the analysis of the
volumes of the segmented Brain, Grey Matter, White Matter and CSF. Then,
functional alterations are studied using statistical parametric DSC MRI
mapping.Il lavoro descritto in questa tesi è stato svolto in collaborazione con l'Unità di
Ricerca in Brain Imaging e Neuropsicologia (Rubin) del Centro Inter
Universitario di Neuroscienze Comportamentali di Udine e Verona (ICBN).
Il gruppo di ricerca studia le alterazioni morfologiche e funzionali del cervello
in pazienti affetti da patologie psichiatriche come, ad esempio, la schizofrenia.
In letteratura sono presenti numerosi articoli sull'argomento, eseguiti per la
maggior parte su immagini ottenute tramite risonanza magnetica strutturale
(MRI) o funzionale (fMRI). Le differenze morfologiche e funzionali tra
pazienti affetti da schizofrenia e controlli sani descritte dalla letteratura sono di
solito molto piccole e caratterizzate da un'elevata variabilità. Questo dipende
dalla variabilità del cervello umano, dall'eventuale uso di farmaci da parte dei
pazienti (che può influenzare le dimensioni e il comportamento dei tessuti) e
dai criteri con cui vengono diagnosticata la schizofrenia (che includono sintomi
profondamente diversi).
La realizzazione di procedure in grado di identificare e analizzare con alta
accuratezza e sensibilità le piccole differenze morfo-funzionali che esistono tra
pazienti affetti da schizofrenia e controlli sani è un problema di elevato
interesse allo stato dell'arte.
Il primo obiettivo del lavoro descritto in questa tesi consiste nell'ideazione, la
realizzazione e la caratterizzazione di un algoritmo totalmente automatico,
robusto, accurato e duttile per la segmentazione del cervello, del fluido cerebro
spinale, della materia grigia e della materia bianca in scansioni MRI T1 pesate
della testa.
Nel campo delle analisi mediche, la segmentazione può essere definita come
l'identificazione dei confini delle diverse strutture anatomiche all'interno delle
immagini. Gli algoritmi di segmentazione sono una componente fondamentale
nella diagnostica per immagini in quanto svolgono un ruolo fondamentale in
numerose applicazioni mediche. Diversi metodi sono stati realizzati per la
segmentazione di specifiche strutture anatomiche.
Per analizzare le piccole alterazioni morfologiche del cervello che sono
correlate alla malattie psichiatriche, sono necessarie procedure di
segmentazione particolarmente sensibili e precise.
Tali procedure devono essere inoltre facilmente modificabili in relazione alle
specifiche esigenze di ricerca. L'Unità di Ricerca in Brain Imaging e
Neuropsicologia dell'ICBN ha deciso di impegnarsi in questo ambito, anziché
utilizzare uno dei software disponibili, perché questi non sono considerati
attualmente soddisfacenti per l'ambito di ricerca e perché desiderava avere uno
strumento proprio, ben conosciuto e facilmente adattabile per specifiche
esigenze. La duttilità è pertanto, una caratteristica fondamentale per l'algoritmo.
Il secondo obiettivo del lavoro è quello di testare la duttilità dell'algoritmo di
segmentazione attraverso la sua applicazione a diverse strutture anatomiche e
tecniche di acquisizione di immagini mediche. Le strutture anatomiche studiate
comprendono il cranio, il cuore, i reni, la vescica, le vie urinarie, lo scheletro e
un tumore del cervello. Le tecniche di acquisizione utilizzate comprendono
scansioni MRI T1 e T2 pesate della testa e una scansione di tomografia
computerizzata angiografica (CTa) del torace. Tali applicazioni possono essere
automatiche o semiautomatiche, a seconda del caso specifico, ma si deve tenere
presente che le procedure semi-automatiche sono ideate in modo da essere
facilmente automatizzate.
Il terzo obiettivo di questo lavoro consiste nell'ideazione, l'implementazione e
la caratterizzazione di una procedura per eseguire voxel per voxel l'analisi
Dynamic Susceptibility Contrast MRI (DSC MRI). La DSC-MRI è una tecnica
di risonanza magnetica di perfusione che ricorre all’uso di un agente di
contrasto esogeno, come il gadolinio, ed è attualmente una delle tecniche più
interessanti per lo studio quantitativo dell’emodinamica cerebrale. La DSC
MRI permette di ricavare importanti parametri emodinamici che ricoprono
un ruolo chiave nello studio di svariate patologie, quali i tumori cerebrali,
l’ischemia, l’infarto, ma studi preliminari suggeriscono che questa tecnica
possa fornire importanti informazioni cliniche anche sui disturbi
neuropsichiatrici (in particolare sulla demenza e la schizofrenia). Allo stato
dell'arte non esistono a nostra conoscenza procedure che permettano di
confrontare voxel per voxel i parametri ottenuti tramite DSC MRI dei pazienti
con quelli dei controlli sani. Questa tecnica potrebbe portare all'individuazione
delle regioni anatomiche maggiormente coinvolte in diverse patologie.
Il quarto obiettivo del lavoro consiste nell'analisi delle alterazioni morfologiche
e funzionali del cervello, del fluido cerebrospinale, della materia grigia e della
materia bianca in pazienti affetti da schizofrenia tramite l'applicazione delle
procedure realizzate.
Le alterazioni morfologiche sono quindi studiate attraverso l'analisi delle
volumetrie delle strutture anatomiche (segmentate tramite l'algoritmo da me
realizzato), mentre le alterazioni funzionali sono studiate utilizzando l'analisi
statistica parametrica voxel per voxel dei dati di DSC MRI
Improving Robotic Hand Prosthesis Control With Eye Tracking and Computer Vision: A Multimodal Approach Based on the Visuomotor Behavior of Grasping
Full text linkISSN:2624-821
Enhancing Gesture Classification Using Active EMG Band and Advanced Feature Extraction Technique
No full textThis study addresses two critical aspects of gesture classification: data acquisition and feature extraction. It introduces an efficient surface electromyogram (sEMG) acquisition band with seven active electrodes and conducts experiments with eight healthy subjects for six gestures. A novel feature extraction method, the Hjorth secant line (HSL), is also introduced. The self-generated dataset (DB1) and two publicly available datasets (DB2 and DB3), containing data from both healthy subjects and amputees, are analyzed to assess the proposed feature set’s performance. The study evaluates the sEMG acquisition system’s effectiveness by measuring the signal-to-noise ratio (SNR), and the performance of the proposed feature set was evaluated in terms of gesture classification accuracy. The feature set’s performance is compared to other existing feature sets in the literature. In addition, a time complexity analysis is performed for the proposed feature set. The SNR for the sEMG acquisition system is 49.36 ± 5.50 dB, demonstrating its efficiency. When using the proposed feature set with a random forest (RF) classifier, the study achieves the classification accuracy of 97.94% ± 0.47%, 98.94% ± 0.46%, and 82.16% ± 2.07% for DB1–DB3, respectively. Paired t -tests indicate that the proposed feature set significantly improves gesture classification accuracy across all three datasets ( p -value < 0.05). The time required to calculate the proposed features was 0.207 ms, which is shorter than the computational time for other feature extraction methods reported in the literature. This study highlights the enhanced SNR from the proposed acquisition system and the feature set’s potential to improve gesture classification. These findings have substantial implications for advancing intelligent prosthetics
Semantic wikis as flexible database interfaces for biomedical applications
Full text linkAbstract Several challenges prevent extracting knowledge from biomedical resources, including data heterogeneity and the difficulty to obtain and collaborate on data and annotations by medical doctors. Therefore, flexibility in their representation and interconnection is required; it is also essential to be able to interact easily with such data. In recent years, semantic tools have been developed: semantic wikis are collections of wiki pages that can be annotated with properties and so combine flexibility and expressiveness, two desirable aspects when modeling databases, especially in the dynamic biomedical domain. However, semantics and collaborative analysis of biomedical data is still an unsolved challenge. The aim of this work is to create a tool for easing the design and the setup of semantic databases and to give the possibility to enrich them with biostatistical applications. As a side effect, this will also make them reproducible, fostering their application by other research groups. A command-line software has been developed for creating all structures required by Semantic MediaWiki. Besides, a way to expose statistical analyses as R Shiny applications in the interface is provided, along with a facility to export Prolog predicates for reasoning with external tools. The developed software allowed to create a set of biomedical databases for the Neuroscience Department of the University of Padova in a more automated way. They can be extended with additional qualitative and statistical analyses of data, including for instance regressions, geographical distribution of diseases, and clustering. The software is released as open source-code and published under the GPL-3 license at https://github.com/mfalda/tsv2swm
Natural control capabilities of robotic hands by hand amputated subjects
No full textPeople with transradial hand amputations who own a myoelectric prosthesis currently have some control capabilities via sEMG. However, the control systems are still limited and not natural. The Ninapro project is aiming at helping the scientific community to overcome these limits through the creation of publicly available electromyography data sources to develop and test machine learning algorithms. In this paper we describe the movement classification results gained from three subjects with an homogeneous level of amputation, and we compare them with the results of 40 intact subjects. The number of considered subjects can seem small at first sight, but it is not considering the literature of the field (which has to face the difficulty of recruiting trans-radial hand amputated subjects). The classification is performed with four different classifiers and the obtained balanced classification rates are up to 58.6% on 50 movements, which is an excellent result compared to the current literature. Successively, for each subject we find a subset of up to 9 highly independent movements, (defined as movements that can be distinguished with more than 90% accuracy), which is a deeply innovative step in literature. The natural control of a robotic hand in so many movements could lead to an immediate progress in robotic hand prosthetics and it could deeply change the quality of life of amputated subjects. © 2014 IEEE
Classification of hand movements in amputated subjects by sEMG and accelerometers
No full textNumerous recent studies have aimed to improve myoelectric control of prostheses. However, the majority of these studies is characterized by two problems that could be easily fulfilled with recent resources supplied by the scientific literature. First, the majority of these studies use only intact subjects, with the unproved assumption that the results apply equally to amputees. Second, usually only electromyography data are used, despite other sensors (e.g., accelerometers) being easy to include into a real life prosthesis control system. In this paper we analyze the mentioned problems by the classification of 40 hand movements in 5 amputated and 40 intact subjects, using both sEMG and accelerometry data and applying several different state of the art methods. The datasets come from the NinaPro database, which supplies publicly available sEMG data to develop and test machine learning algorithms for prosthetics. The number of subjects can seem small at first sight, but it is not considering the literature of the field (which has to face the difficulty of recruiting trans-radial hand amputated subjects). Our results indicate that the maximum average classification accuracy for amputated subjects is 61.14%, which is just 15.86% less than intact subjects, and they show that intact subjects results can be used as proxy measure for amputated subjects. Finally, our comparison shows that accelerometry as a modality is less affected by amputation than electromyography, suggesting that real life prosthetics performance may easily be improved by inclusion of accelerometers. © 2014 IEEE
Quantitative hierarchical representation and comparison of hand grasps from electromyography and kinematic data
Full text linkMotivation: Modeling human grasping and hand movements is important for robotics, prosthetics and rehabilitation. Several qualitative taxonomies of hand grasps have been proposed in scientific literature. However it is not clear how well they correspond to subjects movements.
Objective: In this work we quantitatively analyze the similarity between hand movements in 40 subjects using different features.
Methods: Publicly available data from 40 healthy subjects were used for this study. The data include electromyography and kinematic data
recorded while the subjects perform 20 hand grasps. The kinematic and myoelectric signal was windowed and several signal features were extracted. Then, for each subject, a set of hierarchical trees was computed for the hand grasps. The obtained results were compared in order to evaluate differences between features and different subjects.
Results: The comparison of the signal feature taxonomies revealed a relation among the same subject. The comparison of the subject taxonomies highlighted also a similarity shared between subjects except for rare cases.
Conclusions: The results suggest that quantitative hierarchical representations of hand movements can be performed with the proposed approach and the results from different subjects and features can be compared. The presented approach suggests a way to perform a systematic analysis of hand movements and to create a quantitative taxonomy of hand movements
Automatic classification of canine thoracic radiographs using deep learning.
Full text linkThe interpretation of thoracic radiographs is a challenging and error-prone task for veterinarians. Despite recent advancements in machine learning and computer vision, the development of computer-aided diagnostic systems for radiographs remains a challenging and unsolved problem, particularly in the context of veterinary medicine. In this study, a novel method, based on multi-label deep convolutional neural network (CNN), for the classification of thoracic radiographs in dogs was developed. All the thoracic radiographs of dogs performed between 2010 and 2020 in the institution were retrospectively collected. Radiographs were taken with two different radiograph acquisition systems and were divided into two data sets accordingly. One data set (Data Set 1) was used for training and testing and another data set (Data Set 2) was used to test the generalization ability of the CNNs. Radiographic findings used as non mutually exclusive labels to train the CNNs were: unremarkable, cardiomegaly, alveolar pattern, bronchial pattern, interstitial pattern, mass, pleural effusion, pneumothorax, and megaesophagus. Two different CNNs, based on ResNet-50 and DenseNet-121 architectures respectively, were developed and tested. The CNN based on ResNet-50 had an Area Under the Receive-Operator Curve (AUC) above 0.8 for all the included radiographic findings except for bronchial and interstitial patterns both on Data Set 1 and Data Set 2. The CNN based on DenseNet-121 had a lower overall performance. Statistically significant differences in the generalization ability between the two CNNs were evident, with the CNN based on ResNet-50 showing better performance for alveolar pattern, interstitial pattern, megaesophagus, and pneumothorax
Evaluation of Methods for the Extraction of Spatial Muscle Synergies
Full text linkMuscle synergies have been largely used in many application fields, including motor control studies, prosthesis control, movement classification, rehabilitation, and clinical studies. Due to the complexity of the motor control system, the full repertoire of the underlying synergies has been identified only for some classes of movements and scenarios. Several extraction methods have been used to extract muscle synergies. However, some of these methods may not effectively capture the nonlinear relationship between muscles and impose constraints on input signals or extracted synergies. Moreover, other approaches such as autoencoders (AEs), an unsupervised neural network, were recently introduced to study bioinspired control and movement classification. In this study, we evaluated the performance of five methods for the extraction of spatial muscle synergy, namely, principal component analysis (PCA), independent component analysis (ICA), factor analysis (FA), nonnegative matrix factorization (NMF), and AEs using simulated data and a publicly available database. To analyze the performance of the considered extraction methods with respect to several factors, we generated a comprehensive set of simulated data (ground truth), including spatial synergies and temporal coefficients. The signal-to-noise ratio (SNR) and the number of channels (NoC) varied when generating simulated data to evaluate their effects on ground truth reconstruction. This study also tested the efficacy of each synergy extraction method when coupled with standard classification methods, including K-nearest neighbors (KNN), linear discriminant analysis (LDA), support vector machines (SVM), and Random Forest (RF). The results showed that both SNR and NoC affected the outputs of the muscle synergy analysis. Although AEs showed better performance than FA in variance accounted for and PCA in synergy vector similarity and activation coefficient similarity, NMF and ICA outperformed the other three methods. Classification tasks showed that classification algorithms were sensitive to synergy extraction methods, while KNN and RF outperformed the other two methods for all extraction methods; in general, the classification accuracy of NMF and PCA was higher. Overall, the results suggest selecting suitable methods when performing muscle synergy-related analysis
- …
