Abstract 14131: Proof of Concept: Interpretation of EKG With Image Recognition and Convolutional Neural Networks

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Subrat Das ◽  
Matthew Epland ◽  
Jiang Yu ◽  
Ranjit Suri
Keyword(s):  
Image Recognition ◽  
Gradient Descent ◽  
Confusion Matrix ◽  
Stochastic Gradient Descent ◽  
Training Set ◽  
Data Set ◽  
Computing Power ◽  
The Past ◽  
Novel Approach ◽  
Validation Set

Introduction: EKGs are the cornerstone of management in cardiovascular diseases. There have been multiple efforts to computerize the EKG interpretation with algorithms, which unfortunately are machine specific and proprietary. We propose the development of an image recognition model which can be used to read EKG strips (which use standard notations) and hence be used universally. Method: A convolutional neural network (CNN) was trained to classify 12-lead EKGs between seven clinically important diagnostic classes (Figure 1a). Pre-labeled EKG recordings (6-60s) from a publicly available data set on PhysioNet were used to construct the images. The EKG images displayed the 12 channel traces, of 2.5s each, on a consistent 4x3 grid at a resolution of 800x800 pixels (Figure 1a). The data set (23,336 images) was divided into training, tuning, and validation sets; containing 70%, 15%, and 15% of the images, respectively. An austere variation of the MobileNetV3 model was trained from the ground up on the labeled training set. Stochastic gradient descent (SGD) was used to minimize the cross-entropy loss. Training was halted when the tuning loss had not improved from its previous minimum by 0.05% over the past 10 epochs. Results: The model trained over 52 epochs of batches of 32 images. The model’s accuracy was tested using the validation set (which was not used for development of model) and reported as a confusion matrix (Figure 1b). The accuracy per class varies from 69-91%. Conclusion: We used a labeled dataset of EKG images to develop a CNN model to predict seven different diagnostic classes with good accuracy. This is a novel approach to EKG interpretation as an image recognition problem and thus generates the ability to create diagnostic algorithms that are not dependent on proprietary voltage signals generated by commercial EKG machines. With the addition of more images to the data set and higher computing power we are confident that we can achieve enhanced accuracy.

Performance Analysis of Different Optimizers for Deep Learning-Based Image Recognition

2019 ◽  
Vol 34 (02) ◽  
pp. 2051003 ◽  
Author(s):  
Seda Postalcıoğlu
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Gradient Descent ◽  
Good Accuracy ◽  
Threshold Value ◽  
Training Data ◽  
Stochastic Gradient Descent ◽  
Image Size ◽  
Moment Estimation ◽  
Validation Set

Deep learning refers to Convolutional Neural Network (CNN). CNN is used for image recognition for this study. The dataset is named Fruits-360 and it is obtained from the Kaggle dataset. Seventy percent of the pictures are selected as training data and the rest of the images are used for testing. In this study, an image size is [Formula: see text]. Training is realized using Stochastic Gradient Descent with Momentum (sgdm), Adaptive Moment Estimation (adam) and Root Mean Square Propogation (rmsprop) techniques. The threshold value is determined as 98% for the training. When the accuracy reaches more than 98%, training is stopped. Calculation of the final validation accuracy is done using trained network. In this study, more than 98% of the predicted labels match the true labels of the validation set. Accuracies are calculated using test data for sgdm, adam and rmsprop techniques. The results are 98.08%, 98.85%, 98.88%, respectively. It is clear that fruits are recognized with good accuracy.

Cloning Safe Driving Behavior for Self-Driving Cars using Convolutional Neural Networks

2019 ◽  
Vol 12 (2) ◽  
pp. 120-127 ◽  
Author(s):  
Wael Farag
Keyword(s):  
Gradient Descent ◽  
Autonomous Driving ◽  
Driving Behavior ◽  
Training Data ◽  
Stochastic Gradient Descent ◽  
Data Set ◽  
Safe Driving ◽  
Processing Pipeline ◽  
Self Driving Cars ◽  
And Training

Background: In this paper, a Convolutional Neural Network (CNN) to learn safe driving behavior and smooth steering manoeuvring, is proposed as an empowerment of autonomous driving technologies. The training data is collected from a front-facing camera and the steering commands issued by an experienced driver driving in traffic as well as urban roads. Methods: This data is then used to train the proposed CNN to facilitate what it is called “Behavioral Cloning”. The proposed Behavior Cloning CNN is named as “BCNet”, and its deep seventeen-layer architecture has been selected after extensive trials. The BCNet got trained using Adam’s optimization algorithm as a variant of the Stochastic Gradient Descent (SGD) technique. Results: The paper goes through the development and training process in details and shows the image processing pipeline harnessed in the development. Conclusion: The proposed approach proved successful in cloning the driving behavior embedded in the training data set after extensive simulations.

scholarly journals Parametric UMAP Embeddings for Representation and Semisupervised Learning

2021 ◽  
pp. 1-27
Author(s):  
Tim Sainburg ◽  
Leland McInnes ◽  
Timothy Q. Gentner
Keyword(s):  
Algebraic Topology ◽  
Gradient Descent ◽  
Graphical Representation ◽  
Semisupervised Learning ◽  
Stochastic Gradient Descent ◽  
Second Step ◽  
Data Set ◽  
Parametric Relationship ◽  
Latent Distribution ◽  
Low Dimensional

Abstract UMAP is a nonparametric graph-based dimensionality reduction algorithm using applied Riemannian geometry and algebraic topology to find low-dimensional embeddings of structured data. The UMAP algorithm consists of two steps: (1) computing a graphical representation of a data set (fuzzy simplicial complex) and (2) through stochastic gradient descent, optimizing a low-dimensional embedding of the graph. Here, we extend the second step of UMAP to a parametric optimization over neural network weights, learning a parametric relationship between data and embedding. We first demonstrate that parametric UMAP performs comparably to its nonparametric counterpart while conferring the benefit of a learned parametric mapping (e.g., fast online embeddings for new data). We then explore UMAP as a regularization, constraining the latent distribution of autoencoders, parametrically varying global structure preservation, and improving classifier accuracy for semisupervised learning by capturing structure in unlabeled data.

scholarly journals Application of Multi-Scale Fusion Attention U-Net to Segment the Thyroid Gland on CT Localization Images for Radiotherapy

2021 ◽  
Author(s):  
Xiaobo Wen ◽  
Biao Zhao ◽  
Meifang Yuan ◽  
Jinzhi Li ◽  
Mengzhen Sun ◽  
...  
Keyword(s):  
Thyroid Gland ◽  
Clinical Work ◽  
Similarity Coefficient ◽  
Dice Similarity Coefficient ◽  
Training Set ◽  
Data Set ◽  
Test Set ◽  
Noise Interference ◽  
Multi Scale ◽  
Validation Set

Abstract Objectives: To explore the performance of Multi-scale Fusion Attention U-net (MSFA-U-net) in thyroid gland segmentation on CT localization images for radiotherapy. Methods: CT localization images for radiotherapy of 80 patients with breast cancer or head and neck tumors were selected; label images were manually delineated by experienced radiologists. The data set was randomly divided into the training set (n=60), the validation set (n=10), and the test set (n=10). Data expansion was performed in the training set, and the performance of the MSFA-U-net model was evaluated using the evaluation indicators Dice similarity coefficient (DSC), Jaccard similarity coefficient (JSC), positive predictive value (PPV), sensitivity (SE), and Hausdorff distance (HD). Results: With the MSFA-U-net model, the DSC, JSC, PPV, SE, and HD indexes of the segmented thyroid gland in the test set were 0.8967±0.0935, 0.8219±0.1115, 0.9065±0.0940, 0.8979±0.1104, and 2.3922±0.5423, respectively. Compared with U-net, HR-net, and Attention U-net, MSFA-U-net showed that DSC increased by 0.052, 0.0376, and 0.0346 respectively; JSC increased by 0.0569, 0.0805, and 0.0433, respectively; SE increased by 0.0361, 0.1091, and 0.0831, respectively; and HD increased by −0.208, −0.1952, and −0.0548, respectively. The test set image results showed that the thyroid edges segmented by the MSFA-U-net model were closer to the standard thyroid delineated by the experts, in comparison with those segmented by the other three models. Moreover, the edges were smoother, over-anti-noise interference was stronger, and oversegmentation and undersegmentation were reduced. Conclusion: The MSFA-U-net model can meet basic clinical requirements and improve the efficiency of physicians' clinical work.

A HYBRID SCHEME FOR HANDPRINTED NUMERAL RECOGNITION BASED ON A SELF-ORGANIZING NETWORK AND MLP ClASSIFIERS

2002 ◽  
Vol 16 (07) ◽  
pp. 845-864 ◽  
Author(s):  
UJJWAL BHATTACHARYA ◽  
TANMOY KANTI DAS ◽  
AMITAVA DATTA ◽  
SWAPAN KUMAR PARUI ◽  
BIDYUT BARAN CHAUDHURI
Keyword(s):  
Recognition Rate ◽  
Structural Features ◽  
Data Set ◽  
Novel Approach ◽  
Tree Classifier ◽  
Sample Data ◽  
Validation Set ◽  
Numeral Recognition ◽  
Correct Recognition Rate ◽  
Self Organizing

This paper proposes a novel approach to automatic recognition of handprinted Bangla (an Indian script) numerals. A modified Topology Adaptive Self-Organizing Neural Network is proposed to extract a vector skeleton from a binary numeral image. Simple heuristics are considered to prune artifacts, if any, in such a skeletal shape. Certain topological and structural features like loops, junctions, positions of terminal nodes, etc. are used along with a hierarchical tree classifier to classify handwritten numerals into smaller subgroups. Multilayer perceptron (MLP) networks are then employed to uniquely classify the numerals belonging to each subgroup. The system is trained using a sample data set of 1800 numerals and we have obtained 93.26% correct recognition rate and 1.71% rejection on a separate test set of another 7760 samples. In addition, a validation set consisting of 1440 samples has been used to determine the termination of the training algorithm of the MLP networks. The proposed scheme is sufficiently robust with respect to considerable object noise.

scholarly journals Genome-Wide Identification of a Novel Autophagy-Related Signature for Colorectal Cancer

Dose-Response ◽  
2019 ◽  
Vol 17 (4) ◽  
pp. 155932581989417 ◽  
Author(s):  
Zhi Huang ◽  
Jie Liu ◽  
Liang Luo ◽  
Pan Sheng ◽  
Biao Wang ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Signaling Pathway ◽  
Risk Score ◽  
Low Risk ◽  
Training Data ◽  
The Cancer Genome Atlas ◽  
Training Set ◽  
Data Set ◽  
Validation Set ◽  
Cox Analysis

Background: Plenty of evidence has suggested that autophagy plays a crucial role in the biological processes of cancers. This study aimed to screen autophagy-related genes (ARGs) and establish a novel a scoring system for colorectal cancer (CRC). Methods: Autophagy-related genes sequencing data and the corresponding clinical data of CRC in The Cancer Genome Atlas were used as training data set. The GSE39582 data set from the Gene Expression Omnibus was used as validation set. An autophagy-related signature was developed in training set using univariate Cox analysis followed by stepwise multivariate Cox analysis and assessed in the validation set. Then we analyzed the function and pathways of ARGs using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Finally, a prognostic nomogram combining the autophagy-related risk score and clinicopathological characteristics was developed according to multivariate Cox analysis. Results: After univariate and multivariate analysis, 3 ARGs were used to construct autophagy-related signature. The KEGG pathway analyses showed several significantly enriched oncological signatures, such as p53 signaling pathway, apoptosis, human cytomegalovirus infection, platinum drug resistance, necroptosis, and ErbB signaling pathway. Patients were divided into high- and low-risk groups, and patients with high risk had significantly shorter overall survival (OS) than low-risk patients in both training set and validation set. Furthermore, the nomogram for predicting 3- and 5-year OS was established based on autophagy-based risk score and clinicopathologic factors. The area under the curve and calibration curves indicated that the nomogram showed well accuracy of prediction. Conclusions: Our proposed autophagy-based signature has important prognostic value and may provide a promising tool for the development of personalized therapy.

scholarly journals Deep Convolutional Spiking Neural Networks for Image Classification

2021 ◽  
Author(s):  
Ruthvik Vaila
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Gradient Descent ◽  
Stochastic Gradient ◽  
Spiking Neural Networks ◽  
Stochastic Gradient Descent ◽  
Data Set ◽  
Learning Capabilities ◽  
Artificial Neural

Spiking neural networks are biologically plausible counterparts of artificial neural networks. Artificial neural networks are usually trained with stochastic gradient descent (SGD) and spiking neural networks are trained with bioinspired spike timing dependent plasticity (STDP). Spiking networks could potentially help in reducing power usage owing to their binary activations. In this work, we use unsupervised STDP in the feature extraction layers of a neural network with instantaneous neurons to extract meaningful features. The extracted binary feature vectors are then classified using classification layers containing neurons with binary activations. Gradient descent (backpropagation) is used only on the output layer to perform training for classification. Surrogate gradients are proposed to perform backpropagation with binary gradients. The accuracies obtained for MNIST and the balanced EMNIST data set compare favorably with other approaches. The effect of the stochastic gradient descent (SGD) approximations on learning capabilities of our network are also explored. We also studied catastrophic forgetting and its effect on spiking neural networks (SNNs). For the experiments regarding catastrophic forgetting, in the classification sections of the network we use a modified synaptic intelligence that we refer to as cost per synapse metric as a regularizer to immunize the network against catastrophic forgetting in a Single-Incremental-Task scenario (SIT). In catastrophic forgetting experiments, we use MNIST and EMNIST handwritten digits datasets that were divided into five and ten incremental subtasks respectively. We also examine behavior of the spiking neural network and empirically study the effect of various hyperparameters on its learning capabilities using the software tool SPYKEFLOW that we developed. We employ MNIST, EMNIST and NMNIST data sets to produce our results.

scholarly journals Stochastic gradient descent for hybrid quantum-classical optimization

Quantum ◽  
2020 ◽  
Vol 4 ◽  
pp. 314 ◽  
Author(s):  
Ryan Sweke ◽  
Frederik Wilde ◽  
Johannes Jakob Meyer ◽  
Maria Schuld ◽  
Paul K. Fährmann ◽  
...  
Keyword(s):  
Gradient Descent ◽  
Stochastic Gradient ◽  
Stochastic Gradient Descent ◽  
Expectation Values ◽  
Data Set ◽  
Doubly Stochastic ◽  
Learning Tasks ◽  
Value Estimation ◽  
Near Term ◽  
Classical Optimization

Within the context of hybrid quantum-classical optimization, gradient descent based optimizers typically require the evaluation of expectation values with respect to the outcome of parameterized quantum circuits. In this work, we explore the consequences of the prior observation that estimation of these quantities on quantum hardware results in a form of stochastic gradient descent optimization. We formalize this notion, which allows us to show that in many relevant cases, including VQE, QAOA and certain quantum classifiers, estimating expectation values with k measurement outcomes results in optimization algorithms whose convergence properties can be rigorously well understood, for any value of k. In fact, even using single measurement outcomes for the estimation of expectation values is sufficient. Moreover, in many settings the required gradients can be expressed as linear combinations of expectation values -- originating, e.g., from a sum over local terms of a Hamiltonian, a parameter shift rule, or a sum over data-set instances -- and we show that in these cases k-shot expectation value estimation can be combined with sampling over terms of the linear combination, to obtain ``doubly stochastic'' gradient descent optimizers. For all algorithms we prove convergence guarantees, providing a framework for the derivation of rigorous optimization results in the context of near-term quantum devices. Additionally, we explore numerically these methods on benchmark VQE, QAOA and quantum-enhanced machine learning tasks and show that treating the stochastic settings as hyper-parameters allows for state-of-the-art results with significantly fewer circuit executions and measurements.

Design and Development of Image Recognition Toolkit Based on Deep Learning

2020 ◽  
pp. 2159002
Author(s):  
Hui Zhao ◽  
Hai-Xia Zhang ◽  
Qing-Jiao Cao ◽  
Sheng-Juan Sun ◽  
Xuanzhe Han ◽  
...  
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Gradient Descent ◽  
Recognition Accuracy ◽  
Superior Performance ◽  
Stochastic Gradient Descent ◽  
Experimental Result ◽  
Descent Algorithm ◽  
Learning Platform ◽  
Gradient Descent Algorithm

Deep learning algorithms have shown superior performance than traditional algorithms when dealing with computationally intensive tasks in many fields. The algorithm model based on deep learning has good performance and can improve the recognition accuracy in relevant applications in the field of computer vision. TensorFlow is a flexible opensource machine learning platform proposed by Google, which can run on a variety of platforms, such as CPU, GPU, and mobile devices. TensorFlow platform can also support current popular deep learning models. In this paper, an image recognition toolkit based on TensorFlow is designed and developed to simplify the development process of more and more image recognition applications. The toolkit uses convolutional neural networks to build a training model, which consists of two convolutional layers: one batch normalization layer before each convolutional layer, and the other pooling layer after each convolutional layer. The last two layers of the model use the full connection layer to output recognition results. Batch gradient descent algorithm is adopted in the optimization algorithm, and it integrates the advantages of both the gradient descent algorithm and the stochastic gradient descent algorithm, which greatly reduces the number of convergence iterations and has little influence on the convergence effect. The total training parameters of the toolkit model reach 1.7 million. In order to prevent overfitting problems, the dropout layer before each full connection layer is added and the threshold of 0.5 is set in the design. The convolution neural network model is trained and tested by the MNIST set on TensorFlow. The experimental result shows that the toolkit achieves the recognition accuracy of 99% on the MNIST test set. The development of the toolkit provides powerful technical support for the development of various image recognition applications, reduces its difficulty, and improves the efficiency of resource utilization.

Neural network-based sperm whale click classification

2007 ◽  
Vol 87 (1) ◽  
pp. 35-38 ◽  
Author(s):  
M. van der Schaar ◽  
E. Delory ◽  
A. Català ◽  
M. André
Keyword(s):  
Neural Network ◽  
Wavelet Packet ◽  
Radial Basis Function Network ◽  
Sperm Whale ◽  
Data Sets ◽  
Sperm Whales ◽  
Training Set ◽  
Data Set ◽  
Validation Set ◽  
Function Network

Recordings of a group of foraging sperm whales usually result in a mixture of clicks from different animals. To analyse the click sequences of individual whales these clicks need to be separated, and for this an automatic classifier would be preferred. Here we study the use of a radial basis function network to perform the separation. The neural network's ability to discriminate between different whales was tested with six data sets of individually diving males. The data consisted of five shorter click trains and one complete dive which was especially important to evaluate the capacity of the network to generalize. The network was trained with characteristics extracted from the six click series with the help of a wavelet packet-based local discriminant basis. The selected features were separated in a training set containing 50 clicks of each data set and a validation set with the remaining clicks. After the network was trained it could correctly classify around 90% of the short click series, while for the entire dive this percentage was around 78%.

Sign in / Sign up

Export Citation Format

Share Document