Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635450
N. Filipovic, Christian Helmich, Jasmina Isaković
The brain controls the heart through the sympathetic and parasympathetic branches of the autonomic nervous system. It consists of multisynaptic pathways from myocardial cells back to peripheral ganglionic neurons and further to central preganglionic and premotor neurons. Still, there are no reliable cardiovascular markers of the sympathetic tone and of the sympathetic-parasympathetic balance. It is necessary to understand the interaction between the brain and the heart in order to make early detection and treatment of pathological changes in the brain-heart interaction. In this study we present a detailed electro-chemo-mechanical model of heart and torso, so as to simulate the three principal modes of actions of drugs for cardiomyopathy: (i) modulating calcium transients, (ii) changing kinetics of contractile proteins, (iii) changing the macroscopic structure or its boundary conditions. Heart model geometry included seven different regions. Monodomain model of modified FitzHugh-Nagumo model of the cardiac cell was used. Six electrodes were positioned on the chest to model the precordial leads and the results were compared with real clinical measurements. Inverse ECG method was used to optimize potential on the heart. A whole heart was embedded in the electrical activity throughout the torso environment, with spontaneous initiation of activation in the sinoatrial node, incorporating a specialized conduction system with heterogeneous action potential morphologies throughout the heart. We included body surface potential maps in a healthy subject during progression of ventricular activation in nine sequences. The electrical model was coupled with a mechanical model with orthotropic material properties obtained from the experiments of Holzapfel. In future research we will be more focused on in silico clinical trials with the aim to compare some clinical pathology findings on the body surface with standard 12 ECG electrode measurements.
{"title":"Brain-Heart Electromechanical Modeling","authors":"N. Filipovic, Christian Helmich, Jasmina Isaković","doi":"10.1109/BIBE52308.2021.9635450","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635450","url":null,"abstract":"The brain controls the heart through the sympathetic and parasympathetic branches of the autonomic nervous system. It consists of multisynaptic pathways from myocardial cells back to peripheral ganglionic neurons and further to central preganglionic and premotor neurons. Still, there are no reliable cardiovascular markers of the sympathetic tone and of the sympathetic-parasympathetic balance. It is necessary to understand the interaction between the brain and the heart in order to make early detection and treatment of pathological changes in the brain-heart interaction. In this study we present a detailed electro-chemo-mechanical model of heart and torso, so as to simulate the three principal modes of actions of drugs for cardiomyopathy: (i) modulating calcium transients, (ii) changing kinetics of contractile proteins, (iii) changing the macroscopic structure or its boundary conditions. Heart model geometry included seven different regions. Monodomain model of modified FitzHugh-Nagumo model of the cardiac cell was used. Six electrodes were positioned on the chest to model the precordial leads and the results were compared with real clinical measurements. Inverse ECG method was used to optimize potential on the heart. A whole heart was embedded in the electrical activity throughout the torso environment, with spontaneous initiation of activation in the sinoatrial node, incorporating a specialized conduction system with heterogeneous action potential morphologies throughout the heart. We included body surface potential maps in a healthy subject during progression of ventricular activation in nine sequences. The electrical model was coupled with a mechanical model with orthotropic material properties obtained from the experiments of Holzapfel. In future research we will be more focused on in silico clinical trials with the aim to compare some clinical pathology findings on the body surface with standard 12 ECG electrode measurements.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126722477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635575
Yingtao Wang, Shunfang Wang
The timely identification of plant diseases is crucial for the production of crops. For this problem, many excellent and state-of-the-art algorithms based on deep learning have emerged currently. However, these algorithms still have problems such as poor generalization, difficulty in learning and adapting to new tasks, and extreme reliance on large-scale data. This study introduces an improved meta-learning approach(IMAL) for the few-shot classification of plant diseases, which can produce good generalization performance on new tasks with only a small amount of data and several steps of gradient update. In IMAL, the model-agnostic meta-learning approach with strong generalization capability is used as the overall algorithm framework, a fresh loss function called soft-center loss is adopted to conquer the problem of the poor distinguishing ability of the softmax classifier for features, and the Parametric Rectified Linear Unit (PReLU) activation function is utilized to enhance the model fitting ability with negligible additional computational cost and overfitting risk. The experiment results of plant diseases identification confirmed that the proposed IMAL approach is superior to many current few-shot learning approaches.
{"title":"IMAL: An Improved Meta-learning Approach for Few-shot Classification of Plant Diseases","authors":"Yingtao Wang, Shunfang Wang","doi":"10.1109/BIBE52308.2021.9635575","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635575","url":null,"abstract":"The timely identification of plant diseases is crucial for the production of crops. For this problem, many excellent and state-of-the-art algorithms based on deep learning have emerged currently. However, these algorithms still have problems such as poor generalization, difficulty in learning and adapting to new tasks, and extreme reliance on large-scale data. This study introduces an improved meta-learning approach(IMAL) for the few-shot classification of plant diseases, which can produce good generalization performance on new tasks with only a small amount of data and several steps of gradient update. In IMAL, the model-agnostic meta-learning approach with strong generalization capability is used as the overall algorithm framework, a fresh loss function called soft-center loss is adopted to conquer the problem of the poor distinguishing ability of the softmax classifier for features, and the Parametric Rectified Linear Unit (PReLU) activation function is utilized to enhance the model fitting ability with negligible additional computational cost and overfitting risk. The experiment results of plant diseases identification confirmed that the proposed IMAL approach is superior to many current few-shot learning approaches.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"346 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115411054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635167
Salim Sazzed
Gene expression datasets usually contain a large number of genes which impose a computational burden and complexity on the classifier. Thus, feature selection plays an integral part in sophisticated cancer classification frameworks. In the existing literature, feature selections have been often performed by computationally expensive methods (e.g., wrapper-based methods, evolutionary algorithms). In this paper, we show that the combinations of various simple feature selection methods that require minimal computational cost are often effective for cancer classification. We utilize two sets of simple statistical methods to identify the topmost class-correlated genes (set 1) and eliminate redundant genes (set 2), respectively. Finally, the selected gene set is integrated with the support vector machine (SVM) classifier. The performances of these simple methodologies are compared with a number of existing methods on ten gene expression benchmark datasets. It is observed that in many datasets, these simple methodologies yield similar efficacy to the complex and computationally expensive approaches using only a small number of genes.
{"title":"An investigation of the performances of simple gene selection methodologies for cancer classification","authors":"Salim Sazzed","doi":"10.1109/BIBE52308.2021.9635167","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635167","url":null,"abstract":"Gene expression datasets usually contain a large number of genes which impose a computational burden and complexity on the classifier. Thus, feature selection plays an integral part in sophisticated cancer classification frameworks. In the existing literature, feature selections have been often performed by computationally expensive methods (e.g., wrapper-based methods, evolutionary algorithms). In this paper, we show that the combinations of various simple feature selection methods that require minimal computational cost are often effective for cancer classification. We utilize two sets of simple statistical methods to identify the topmost class-correlated genes (set 1) and eliminate redundant genes (set 2), respectively. Finally, the selected gene set is integrated with the support vector machine (SVM) classifier. The performances of these simple methodologies are compared with a number of existing methods on ten gene expression benchmark datasets. It is observed that in many datasets, these simple methodologies yield similar efficacy to the complex and computationally expensive approaches using only a small number of genes.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115694174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635238
Oguzhan Kirtas, M. Mohammadi, B. Bentsen, P. Veltink, L. Struijk
Tongue-computer interfaces have shown the potential to control assistive devices developed for individuals with severe disabilities. However, current efficient tongue-computer interfaces require invasive methods for attaching the sensor activation units to the tongue, such as piercing. In this study, we propose a noninvasive tongue-computer interface to avoid the requirement of invasive activation unit attachment methods. We developed the noninvasive tongue-computer interface by integrating an activation unit on a frame, and mounting the frame on an inductive tongue-computer interface (ITCI). Thus, the users are able to activate the inductive sensors on the interface by positioning the activation unit with their tongue. They also do not need to remount the activation unit before each use. We performed pointing tests for controlling a computer cursor and number typing tests with two able-bodied participants, where one of them was experienced with using invasive tongue-computer interfaces and other one had no experience. We measured throughput and movement error for pointing tasks, and speed and accuracy for number typing tasks for the evaluation of the feasibility and performance of the developed noninvasive system. Results show that the inexperienced participant achieved similar results with the developed noninvasive tongue-computer interface compared to the current invasive version of the ITCI, while the experienced participant performed better with the invasive tongue-computer interface.
{"title":"Design and evaluation of a noninvasive tongue-computer interface for individuals with severe disabilities","authors":"Oguzhan Kirtas, M. Mohammadi, B. Bentsen, P. Veltink, L. Struijk","doi":"10.1109/BIBE52308.2021.9635238","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635238","url":null,"abstract":"Tongue-computer interfaces have shown the potential to control assistive devices developed for individuals with severe disabilities. However, current efficient tongue-computer interfaces require invasive methods for attaching the sensor activation units to the tongue, such as piercing. In this study, we propose a noninvasive tongue-computer interface to avoid the requirement of invasive activation unit attachment methods. We developed the noninvasive tongue-computer interface by integrating an activation unit on a frame, and mounting the frame on an inductive tongue-computer interface (ITCI). Thus, the users are able to activate the inductive sensors on the interface by positioning the activation unit with their tongue. They also do not need to remount the activation unit before each use. We performed pointing tests for controlling a computer cursor and number typing tests with two able-bodied participants, where one of them was experienced with using invasive tongue-computer interfaces and other one had no experience. We measured throughput and movement error for pointing tasks, and speed and accuracy for number typing tasks for the evaluation of the feasibility and performance of the developed noninvasive system. Results show that the inexperienced participant achieved similar results with the developed noninvasive tongue-computer interface compared to the current invasive version of the ITCI, while the experienced participant performed better with the invasive tongue-computer interface.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114517045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635402
Momcilo Prodanovic, Danica Prodanovic, B. Stojanovic, N. Filipovic, Gordana R. Jovicic, S. Mijailovich
Continuous degeneration of muscle tissue, inflammatory processes and fibrosis characterized by a loss of muscle mass, formation of micro-scars, adipose tissue in the muscles and eventual muscle punctures are often signs of muscular dystrophies (dystrophinopathies). These neuromuscular diseases result from genetic mutations of a structural protein called dystrophin. The absence of functional dystrophin leads to the most common and severe form of muscular dystrophy, Duchenne muscular dystrophy (DMD). Typically, within one muscle bundle there are so-called fast and slow muscle fibers that shorten and lengthen at different speeds during muscle contraction. Using the multiscale muscle platform Mexie we evaluated how the lack of dystrophin affects the connective tissue deformation between these two types of muscle fibers. By adjusting the elasticity of extracellular matrix layer, we estimated the magnitude of the shear strain under unloaded and lightly loaded fiber contractions caused by differences in shortening velocities between fast and slow fibers. The simulations showed that without dystrophin large shear strains are generated causing local micro injury and inflammation leading to further muscle degeneration. The multiscale muscle modeling approach presented here could help accelerate understanding of DMD and lead to faster development of new drugs and treatments of patients.
{"title":"Estimation of Shear Stress Variation in Extracellular Matrix Caused by Duchenne Muscular Dystrophy","authors":"Momcilo Prodanovic, Danica Prodanovic, B. Stojanovic, N. Filipovic, Gordana R. Jovicic, S. Mijailovich","doi":"10.1109/BIBE52308.2021.9635402","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635402","url":null,"abstract":"Continuous degeneration of muscle tissue, inflammatory processes and fibrosis characterized by a loss of muscle mass, formation of micro-scars, adipose tissue in the muscles and eventual muscle punctures are often signs of muscular dystrophies (dystrophinopathies). These neuromuscular diseases result from genetic mutations of a structural protein called dystrophin. The absence of functional dystrophin leads to the most common and severe form of muscular dystrophy, Duchenne muscular dystrophy (DMD). Typically, within one muscle bundle there are so-called fast and slow muscle fibers that shorten and lengthen at different speeds during muscle contraction. Using the multiscale muscle platform Mexie we evaluated how the lack of dystrophin affects the connective tissue deformation between these two types of muscle fibers. By adjusting the elasticity of extracellular matrix layer, we estimated the magnitude of the shear strain under unloaded and lightly loaded fiber contractions caused by differences in shortening velocities between fast and slow fibers. The simulations showed that without dystrophin large shear strains are generated causing local micro injury and inflammation leading to further muscle degeneration. The multiscale muscle modeling approach presented here could help accelerate understanding of DMD and lead to faster development of new drugs and treatments of patients.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130775984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635175
Lin Li, Wonseok Seo
According to Skin Cancer Foundation, skin cancer is by far the most common type of cancer in the United States and worldwide. Early diagnosis of skin cancer is critical because proper treatment at early stages can increase the chance of cure and recovery. However, visual inspection of dermoscopic images by dermatologists is error-prone and time-consuming. To ensure accurate diagnosis and faster treatment of skin cancer, deep learning techniques have been utilized to conduct automated skin lesion segmentation and classification. In this paper, after image processing, a Mask R-CNN model is built for lesion segmentation, where transfer learning is utilized by using the pre-trained weights from Microsoft COCO dataset. The weights of the trained Mask R-CNN model are saved and transferred to the next task - skin lesion classification, to train a Mask R-CNN model for classification. Our experiments are conducted on the benchmark datasets from the International Skin Imaging Collaboration 2018 (ISIC 2018) and evaluated by the same metrics used in ISIC 2018. The lesion boundary segmentation and lesion classification have achieved an accuracy of 96% and a balanced multiclass accuracy of 80%, respectively.
{"title":"Deep Learning and Transfer Learning for Skin Cancer Segmentation and Classification","authors":"Lin Li, Wonseok Seo","doi":"10.1109/BIBE52308.2021.9635175","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635175","url":null,"abstract":"According to Skin Cancer Foundation, skin cancer is by far the most common type of cancer in the United States and worldwide. Early diagnosis of skin cancer is critical because proper treatment at early stages can increase the chance of cure and recovery. However, visual inspection of dermoscopic images by dermatologists is error-prone and time-consuming. To ensure accurate diagnosis and faster treatment of skin cancer, deep learning techniques have been utilized to conduct automated skin lesion segmentation and classification. In this paper, after image processing, a Mask R-CNN model is built for lesion segmentation, where transfer learning is utilized by using the pre-trained weights from Microsoft COCO dataset. The weights of the trained Mask R-CNN model are saved and transferred to the next task - skin lesion classification, to train a Mask R-CNN model for classification. Our experiments are conducted on the benchmark datasets from the International Skin Imaging Collaboration 2018 (ISIC 2018) and evaluated by the same metrics used in ISIC 2018. The lesion boundary segmentation and lesion classification have achieved an accuracy of 96% and a balanced multiclass accuracy of 80%, respectively.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123052178","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635303
R. L. Kæseler, L. Struijk, M. Jochumsen
While assistive robotic devices can improve the quality of life for individuals with tetraplegia, it is difficult to provide a high-performing interface that can be fully utilized, with little to no motor functionality. While a brain-computer interface (BCI) can be used with little to no motor functionality, it typically has a low performance. Steady-state visually evoked potentials (SSVEP) provide some of the best performing signals for a BCI, but are rarely investigated for online asynchronous control where not only accuracy is important, but also the computational costs. This study investigates and compares three classifiers: the well-known and high-performing task-related component analysis (TRCA), the computational efficient Spatiotemporal beamformer (STBF) build on the stimulus-locked inter-trace correlation (SLIC) algorithm and our proposed novel algorithm which combines the two: the SLIC-TRCA. Results show the SLIC-TRCA achieving higher accuracies ${(95.00pm 5.36%}$ with a 1s classification window) compared to the TRCA ${(88.25pm 14.58%)}$ and similar compared to the STBF ${(91.00pm 11.02%)}$ while having a much lower computational cost (519% faster than the TRCA and 144% faster than the STBF). We, therefore, believe this algorithm has an exciting potential as it will allow a high classification accuracy without requiring a high-performing CPU.
{"title":"Optimizing steady-state visual evoked potential classifiers for high performance and low computational costs in brain-computer interfacing","authors":"R. L. Kæseler, L. Struijk, M. Jochumsen","doi":"10.1109/BIBE52308.2021.9635303","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635303","url":null,"abstract":"While assistive robotic devices can improve the quality of life for individuals with tetraplegia, it is difficult to provide a high-performing interface that can be fully utilized, with little to no motor functionality. While a brain-computer interface (BCI) can be used with little to no motor functionality, it typically has a low performance. Steady-state visually evoked potentials (SSVEP) provide some of the best performing signals for a BCI, but are rarely investigated for online asynchronous control where not only accuracy is important, but also the computational costs. This study investigates and compares three classifiers: the well-known and high-performing task-related component analysis (TRCA), the computational efficient Spatiotemporal beamformer (STBF) build on the stimulus-locked inter-trace correlation (SLIC) algorithm and our proposed novel algorithm which combines the two: the SLIC-TRCA. Results show the SLIC-TRCA achieving higher accuracies ${(95.00pm 5.36%}$ with a 1s classification window) compared to the TRCA ${(88.25pm 14.58%)}$ and similar compared to the STBF ${(91.00pm 11.02%)}$ while having a much lower computational cost (519% faster than the TRCA and 144% faster than the STBF). We, therefore, believe this algorithm has an exciting potential as it will allow a high classification accuracy without requiring a high-performing CPU.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116977730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635319
Aleksandra Vulovic, G. Filardo, N. Filipovic
During everyday activities cartilage experiences high loads, stresses, deformations, and contact forces. Sometimes, those activities can lead to permanent damage, such as focal lesions. Focal cartilage lesions have been associated with the progressive degeneration of the surrounding cartilage tissue. This paper aims to compare the mechanical response of the knee joint and femoral cartilage using finite element models during the stance phase of the gait cycle. Our model, developed from MRI scans, has been used to compare the mechanical response of the knee joint with healthy and damaged femoral cartilage. The location of the lesion was above the anterior section of the lateral meniscus. Comparison of the obtained results has shown that having a lesion in the previously mentioned location leads to a significantly higher peak Von Mises stress values.
{"title":"Comparison of mechanical response of knee joint with healthy and damaged femoral cartilage","authors":"Aleksandra Vulovic, G. Filardo, N. Filipovic","doi":"10.1109/BIBE52308.2021.9635319","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635319","url":null,"abstract":"During everyday activities cartilage experiences high loads, stresses, deformations, and contact forces. Sometimes, those activities can lead to permanent damage, such as focal lesions. Focal cartilage lesions have been associated with the progressive degeneration of the surrounding cartilage tissue. This paper aims to compare the mechanical response of the knee joint and femoral cartilage using finite element models during the stance phase of the gait cycle. Our model, developed from MRI scans, has been used to compare the mechanical response of the knee joint with healthy and damaged femoral cartilage. The location of the lesion was above the anterior section of the lateral meniscus. Comparison of the obtained results has shown that having a lesion in the previously mentioned location leads to a significantly higher peak Von Mises stress values.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"108 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133828852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635261
J. Musulin, D. Štifanić, Ana Zulijani, Sandi Baressi Segota, I. Lorencin, N. Anđelić, Z. Car
The diagnosis of oral squamous cell carcinoma is based on a histopathological examination, which is still the most reliable way of identifying oral cancer despite its high subjectivity. However, due to the heterogeneous structure and textures of oral cancer, as well as the presence of any inflammatory tissue reaction, histopathological classification can be difficult. For that reason, an automatic classification of histopathology images with the help of artificial intelligence-assisted technologies can not only improve objective diagnostic results for the clinician but also provide extensive texture analysis to get a correct diagnosis. In this paper various deep learning methods are compared in order to get an AI-based model for multiclass grading of OSCC with the highest $mathbf{AUC}_{mathbf{micro}}$ and ${text{AUC}}_{text{macro}}$ values.
{"title":"Automated Grading of Oral Squamous Cell Carcinoma into Multiple Classes Using Deep Learning Methods","authors":"J. Musulin, D. Štifanić, Ana Zulijani, Sandi Baressi Segota, I. Lorencin, N. Anđelić, Z. Car","doi":"10.1109/BIBE52308.2021.9635261","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635261","url":null,"abstract":"The diagnosis of oral squamous cell carcinoma is based on a histopathological examination, which is still the most reliable way of identifying oral cancer despite its high subjectivity. However, due to the heterogeneous structure and textures of oral cancer, as well as the presence of any inflammatory tissue reaction, histopathological classification can be difficult. For that reason, an automatic classification of histopathology images with the help of artificial intelligence-assisted technologies can not only improve objective diagnostic results for the clinician but also provide extensive texture analysis to get a correct diagnosis. In this paper various deep learning methods are compared in order to get an AI-based model for multiclass grading of OSCC with the highest $mathbf{AUC}_{mathbf{micro}}$ and ${text{AUC}}_{text{macro}}$ values.","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132092795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-10-25DOI: 10.1109/BIBE52308.2021.9635330
P. Priya, Srinivasan Jayaraman
Factors inducing Hydroxychloroquine (HCQ) car-diotoxicity are still unclear, and this paper attempts to understand whether the presence of hypoxia in a congenital long QT syndrome1 (LQTS1) ventricular tissue can affect the outcome of HCQ interaction. This is facilitated by analysing the combination of LQTS1, HCQ and, mild and severe hypoxic conditions in a) the three types of cardiomyocytes: endocardial, midmyocardial and epicardial, as well as b) by generating pseudo ECGs from a 2D transmural anisotropic ventricular tissue model that has been excited with premature beats (PBs) to understand the possibility of arrhythmic occurrence. Results show that inclusion of HCQ in LQTS1 conditions prolongs the action potential duration(APD) in all cell types, leading to early after depolarisations (EADs) in M-cells alone. In contrast, on including hypoxia, the APDs are shortened in all cell types. Pseudo ECGs show a QT interval prolongation on adding HCQ with LQTS1 condition. In addition to LQTS1, mild and severe hypoxia, induces QT interval reduction, with low amplitude notched or inverted T-wave respectively. In presence of PBs, premature ventricular complexes (PVCs) are generated only in presence of HCQ with LQTS1. However, no significant effect of HCQ is observed in both hypoxia severities. Clinical relevance-This in-silico ventricular model indicates that although LQTS1 patients might be contraindicated for HCQ treatment, the combination of mild hypoxia and LQTS1 doesn't pose a risk factor and could help guide HCQ therapy
{"title":"Does the influence of hydroxychloroquine in a hypoxic ventricle differ from that of a non-hypoxic ventricle under congenital LQTS1 ?","authors":"P. Priya, Srinivasan Jayaraman","doi":"10.1109/BIBE52308.2021.9635330","DOIUrl":"https://doi.org/10.1109/BIBE52308.2021.9635330","url":null,"abstract":"Factors inducing Hydroxychloroquine (HCQ) car-diotoxicity are still unclear, and this paper attempts to understand whether the presence of hypoxia in a congenital long QT syndrome1 (LQTS1) ventricular tissue can affect the outcome of HCQ interaction. This is facilitated by analysing the combination of LQTS1, HCQ and, mild and severe hypoxic conditions in a) the three types of cardiomyocytes: endocardial, midmyocardial and epicardial, as well as b) by generating pseudo ECGs from a 2D transmural anisotropic ventricular tissue model that has been excited with premature beats (PBs) to understand the possibility of arrhythmic occurrence. Results show that inclusion of HCQ in LQTS1 conditions prolongs the action potential duration(APD) in all cell types, leading to early after depolarisations (EADs) in M-cells alone. In contrast, on including hypoxia, the APDs are shortened in all cell types. Pseudo ECGs show a QT interval prolongation on adding HCQ with LQTS1 condition. In addition to LQTS1, mild and severe hypoxia, induces QT interval reduction, with low amplitude notched or inverted T-wave respectively. In presence of PBs, premature ventricular complexes (PVCs) are generated only in presence of HCQ with LQTS1. However, no significant effect of HCQ is observed in both hypoxia severities. Clinical relevance-This in-silico ventricular model indicates that although LQTS1 patients might be contraindicated for HCQ treatment, the combination of mild hypoxia and LQTS1 doesn't pose a risk factor and could help guide HCQ therapy","PeriodicalId":343724,"journal":{"name":"2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132820498","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：