scholarly journals A High Accuracy Electrographic Seizure Classifier Trained Using Semi-Supervised Labeling Applied to a Large Spectrogram Dataset

2021 ◽  
Vol 15 ◽  
Author(s):  
Wade Barry ◽  
Sharanya Arcot Desai ◽  
Thomas K. Tcheng ◽  
Martha J. Morrell
Keyword(s):  
Classification Accuracy ◽  
Training Data ◽  
Model Complexity ◽  
Image Clustering ◽  
Test Accuracy ◽  
Patient Classification ◽  
Electrographic Seizure ◽  
Number Of Patients ◽  
Independent Expert ◽  
Model Training

The objective of this study was to explore using ECoG spectrogram images for training reliable cross-patient electrographic seizure classifiers, and to characterize the classifiers’ test accuracy as a function of amount of training data. ECoG channels in ∼138,000 time-series ECoG records from 113 patients were converted to RGB spectrogram images. Using an unsupervised spectrogram image clustering technique, manual labeling of 138,000 ECoG records (each with up to 4 ECoG channels) was completed in 320 h, which is an estimated 5 times faster than manual labeling without ECoG clustering. For training supervised classifier models, five random folds of data were created; with each fold containing 72, 18, and 23 patients’ data for model training, validation and testing respectively. Five convolutional neural network (CNN) architectures, including two with residual connections, were trained. Cross-patient classification accuracies and F1 scores improved with model complexity, with the shallowest 6-layer model (with ∼1.5 million trainable parameters) producing a class-balanced seizure/non-seizure classification accuracy of 87.9% on ECoG channels and the deepest ResNet50-based model (with ∼23.5 million trainable parameters) producing a classification accuracy of 95.7%. The trained ResNet50-based model additionally had 93.5% agreement in scores with an independent expert labeller. Visual inspection of gradient-based saliency maps confirmed that the models’ classifications were based on relevant portions of the spectrogram images. Further, by repeating training experiments with data from varying number of patients, it was found that ECoG spectrogram images from just 10 patients were sufficient to train ResNet50-based models with 88% cross-patient accuracy, while at least 30 patients’ data was required to produce cross-patient classification accuracies of >90%.

scholarly journals SEMI-SUPERVISED SEQUENCE CLASSIFICATION WITH HMMs

2005 ◽  
Vol 19 (02) ◽  
pp. 165-182 ◽  
Author(s):  
SHI ZHONG
Keyword(s):  
Supervised Learning ◽  
Learning Strategies ◽  
Test Data ◽  
Unlabeled Data ◽  
Training Data ◽  
Model Complexity ◽  
Model Parameters ◽  
Training Process ◽  
Transductive Learning ◽  
Model Training

Using unlabeled data to help supervised learning has become an increasingly attractive methodology and proven to be effective in many applications. This paper applies semi-supervised classification algorithms, based on hidden Markov models, to classify sequences. For model-based classification, semi-supervised learning amounts to using both labeled and unlabeled data to train model parameters. We examine three different strategies of using labeled and unlabeled data in the model training process. These strategies differ in how and when labeled and unlabeled data contribute to the model training process. We also compare regular semi-supervised learning, where there are separate unlabeled training data and unlabeled test data, with transductive learning where we do not differentiate between unlabeled training data and unlabeled test data. Our experimental results on synthetic and real EEG time-series show that substantially improved classification accuracy can be achieved by these semi-supervised learning strategies. The effect of model complexity on semi-supervised learning is also studied in our experiments.

Abductive Network Committees for Improved Classification of Medical Data

2004 ◽  
Vol 43 (02) ◽  
pp. 192-201 ◽  
Author(s):  
R. E. Abdel-Aal
Keyword(s):  
Classification Accuracy ◽  
Network Models ◽  
Training Data ◽  
Model Complexity ◽  
Pima Indians ◽  
Training Set ◽  
Combination Methods ◽  
Speed Up ◽  
Abductive Network

Summary Objectives: To introduce abductive network classifier committees as an ensemble method for improving classification accuracy in medical diagnosis. While neural networks allow many ways to introduce enough diversity among member models to improve performance when forming a committee, the self-organizing, automatic-stopping nature, and learning approach used by abductive networks are not very conducive for this purpose. We explore ways of overcoming this limitation and demonstrate improved classification on three standard medical datasets. Methods: Two standard 2-class medical datasets (Pima Indians Diabetes and Heart Disease) and a 6-class dataset (Dermatology) were used to investigate ways of training abductive networks with adequate independence, as well as methods of combining their outputs to form a network that improves performance beyond that of single models. Results: Two- or three-member committees of models trained on completely or partially different subsets of training data and using simple output combination methods achieve improvements between 2 and 5 percentage points in the classification accuracy over the best single model developed using the full training set. Conclusions: Varying model complexity alone gives abductive network models that are too correlated to ensure enough diversity for forming a useful committee. Diversity achieved through training member networks on independent subsets of the training data outweighs limitations of the smaller training set for each, resulting in net gain in committee performance. As such models train faster and can be trained in parallel, this can also speed up classifier development.

Additive DenseNet: Dense connections based on simple addition operations

2020 ◽  
pp. 1-11
Author(s):  
Dawei Yu ◽  
Jie Yang ◽  
Yun Zhang ◽  
Shujuan Yu
Keyword(s):  
Text Classification ◽  
Deep Neural Networks ◽  
Model Complexity ◽  
Residual Network ◽  
Large Margin ◽  
Simple Addition ◽  
High Resource ◽  
Model Training ◽  
Low Efficiency ◽  
Resource Economy

The Densely Connected Network (DenseNet) has been widely recognized as a highly competitive architecture in Deep Neural Networks. And its most outstanding property is called Dense Connections, which represent each layer’s input by concatenating all the preceding layers’ outputs and thus improve the performance by encouraging feature reuse to the extreme. However, it is Dense Connections that cause the challenge of dimension-enlarging, making DenseNet very resource-intensive and low efficiency. In the light of this, inspired by the Residual Network (ResNet), we propose an improved DenseNet named Additive DenseNet, which features replacing concatenation operations (used in Dense Connections) with addition operations (used in ResNet), and in terms of feature reuse, it upgrades addition operations to accumulating operations (namely ∑ (·)), thus enables each layer’s input to be the summation of all the preceding layers’ outputs. Consequently, Additive DenseNet can not only preserve the dimension of input from enlarging, but also retain the effect of Dense Connections. In this paper, Additive DenseNet is applied to text classification task. The experimental results reveal that compared to DenseNet, our Additive DenseNet can reduce the model complexity by a large margin, such as GPU memory usage and quantity of parameters. And despite its high resource economy, Additive DenseNet can still outperform DenseNet on 6 text classification datasets in terms of accuracy and show competitive performance for model training.

Fully automated contrast and non-contrast cardiac view detection in echocardiography a multi-centre, multi-vendor study

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
S Gao ◽  
D Stojanovski ◽  
A Parker ◽  
P Marques ◽  
S Heitner ◽  
...  
Keyword(s):  
Neural Network ◽  
Training Data ◽  
Classification Model ◽  
Validation Dataset ◽  
Funding Source ◽  
Private Company ◽  
Validation Data ◽  
Independent Test ◽  
Model Training ◽  
Confusion Matrices

Abstract Background Correctly identifying views acquired in a 2D echocardiographic examination is paramount to post-processing and quantification steps often performed as part of most clinical workflows. In many exams, particularly in stress echocardiography, microbubble contrast is used which greatly affects the appearance of the cardiac views. Here we present a bespoke, fully automated convolutional neural network (CNN) which identifies apical 2, 3, and 4 chamber, and short axis (SAX) views acquired with and without contrast. The CNN was tested in a completely independent, external dataset with the data acquired in a different country than that used to train the neural network. Methods Training data comprised of 2D echocardiograms was taken from 1014 subjects from a prospective multisite, multi-vendor, UK trial with the number of frames in each view greater than 17,500. Prior to view classification model training, images were processed using standard techniques to ensure homogenous and normalised image inputs to the training pipeline. A bespoke CNN was built using the minimum number of convolutional layers required with batch normalisation, and including dropout for reducing overfitting. Before processing, the data was split into 90% for model training (211,958 frames), and 10% used as a validation dataset (23,946 frames). Image frames from different subjects were separated out entirely amongst the training and validation datasets. Further, a separate trial dataset of 240 studies acquired in the USA was used as an independent test dataset (39,401 frames). Results Figure 1 shows the confusion matrices for both validation data (left) and independent test data (right), with an overall accuracy of 96% and 95% for the validation and test datasets respectively. The accuracy for the non-contrast cardiac views of >99% exceeds that seen in other works. The combined datasets included images acquired across ultrasound manufacturers and models from 12 clinical sites. Conclusion We have developed a CNN capable of automatically accurately identifying all relevant cardiac views used in “real world” echo exams, including views acquired with contrast. Use of the CNN in a routine clinical workflow could improve efficiency of quantification steps performed after image acquisition. This was tested on an independent dataset acquired in a different country to that used to train the model and was found to perform similarly thus indicating the generalisability of the model. Figure 1. Confusion matrices Funding Acknowledgement Type of funding source: Private company. Main funding source(s): Ultromics Ltd.

scholarly journals BeautyNet: Joint Multiscale CNN and Transfer Learning Method for Unconstrained Facial Beauty Prediction

2019 ◽  
Vol 2019 ◽  
pp. 1-14 ◽  
Author(s):  
Yikui Zhai ◽  
He Cao ◽  
Wenbo Deng ◽  
Junying Gan ◽  
Vincenzo Piuri ◽  
...  
Keyword(s):  
Transfer Learning ◽  
Classification Accuracy ◽  
Learning Strategy ◽  
State Of The Art ◽  
Activation Function ◽  
Training Data ◽  
Fine Grained ◽  
Pattern Recognition Problem ◽  
Face Features ◽  
Facial Beauty

Because of the lack of discriminative face representations and scarcity of labeled training data, facial beauty prediction (FBP), which aims at assessing facial attractiveness automatically, has become a challenging pattern recognition problem. Inspired by recent promising work on fine-grained image classification using the multiscale architecture to extend the diversity of deep features, BeautyNet for unconstrained facial beauty prediction is proposed in this paper. Firstly, a multiscale network is adopted to improve the discriminative of face features. Secondly, to alleviate the computational burden of the multiscale architecture, MFM (max-feature-map) is utilized as an activation function which can not only lighten the network and speed network convergence but also benefit the performance. Finally, transfer learning strategy is introduced here to mitigate the overfitting phenomenon which is caused by the scarcity of labeled facial beauty samples and improves the proposed BeautyNet’s performance. Extensive experiments performed on LSFBD demonstrate that the proposed scheme outperforms the state-of-the-art methods, which can achieve 67.48% classification accuracy.

scholarly journals ChemProps: A RESTful API enabled database for composite polymer name standardization

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Bingyin Hu ◽  
Anqi Lin ◽  
L. Catherine Brinson
Keyword(s):  
Data Exchange ◽  
Ground Truth ◽  
Training Data ◽  
Test Accuracy ◽  
Polymer Science ◽  
Data Systems ◽  
Materials Informatics ◽  
Related Data ◽  
Weight Factors ◽  
Data Points

AbstractThe inconsistency of polymer indexing caused by the lack of uniformity in expression of polymer names is a major challenge for widespread use of polymer related data resources and limits broad application of materials informatics for innovation in broad classes of polymer science and polymeric based materials. The current solution of using a variety of different chemical identifiers has proven insufficient to address the challenge and is not intuitive for researchers. This work proposes a multi-algorithm-based mapping methodology entitled ChemProps that is optimized to solve the polymer indexing issue with easy-to-update design both in depth and in width. RESTful API is enabled for lightweight data exchange and easy integration across data systems. A weight factor is assigned to each algorithm to generate scores for candidate chemical names and optimized to maximize the minimum value of the score difference between the ground truth chemical name and the other candidate chemical names. Ten-fold validation is utilized on the 160 training data points to prevent overfitting issues. The obtained set of weight factors achieves a 100% test accuracy on the 54 test data points. The weight factors will evolve as ChemProps grows. With ChemProps, other polymer databases can remove duplicate entries and enable a more accurate “search by SMILES” function by using ChemProps as a common name-to-SMILES translator through API calls. ChemProps is also an excellent tool for auto-populating polymer properties thanks to its easy-to-update design.

Deep Learning based Tomato’s Ripe and Unripe Classification System

2022 ◽  
Vol 10 (1) ◽  
pp. 0-0
Keyword(s):  
Deep Learning ◽  
Classification Accuracy ◽  
Ccd Camera ◽  
Agricultural Products ◽  
Training Data ◽  
Maturity Level ◽  
Agriculture Sector ◽  
The Past ◽  
State Of Art

Effective productivity estimates of fresh produced crops are very essential for efficient farming, commercial planning, and logistical support. In the past ten years, machine learning (ML) algorithms have been widely used for grading and classification of agricultural products in agriculture sector. However, the precise and accurate assessment of the maturity level of tomatoes using ML algorithms is still a quite challenging to achieve due to these algorithms being reliant on hand crafted features. Hence, in this paper we propose a deep learning based tomato maturity grading system that helps to increase the accuracy and adaptability of maturity grading tasks with less amount of training data. The performance of proposed system is assessed on the real tomato datasets collected from the open fields using Nikon D3500 CCD camera. The proposed approach achieved an average maturity classification accuracy of 99.8 % which seems to be quite promising in comparison to the other state of art methods.

scholarly journals Medical Text Classification Using Hybrid Deep Learning Models with Multihead Attention

2021 ◽  
Vol 2021 ◽  
pp. 1-16
Author(s):  
Sunil Kumar Prabhakar ◽  
Dong-Ok Won
Keyword(s):  
Deep Learning ◽  
Language Processing ◽  
Text Classification ◽  
Patient Information ◽  
Classification Accuracy ◽  
Learning Model ◽  
Training Data ◽  
Machine Learning Techniques ◽  
Medical Text ◽  
Deep Learning Model

To unlock information present in clinical description, automatic medical text classification is highly useful in the arena of natural language processing (NLP). For medical text classification tasks, machine learning techniques seem to be quite effective; however, it requires extensive effort from human side, so that the labeled training data can be created. For clinical and translational research, a huge quantity of detailed patient information, such as disease status, lab tests, medication history, side effects, and treatment outcomes, has been collected in an electronic format, and it serves as a valuable data source for further analysis. Therefore, a huge quantity of detailed patient information is present in the medical text, and it is quite a huge challenge to process it efficiently. In this work, a medical text classification paradigm, using two novel deep learning architectures, is proposed to mitigate the human efforts. The first approach is that a quad channel hybrid long short-term memory (QC-LSTM) deep learning model is implemented utilizing four channels, and the second approach is that a hybrid bidirectional gated recurrent unit (BiGRU) deep learning model with multihead attention is developed and implemented successfully. The proposed methodology is validated on two medical text datasets, and a comprehensive analysis is conducted. The best results in terms of classification accuracy of 96.72% is obtained with the proposed QC-LSTM deep learning model, and a classification accuracy of 95.76% is obtained with the proposed hybrid BiGRU deep learning model.

scholarly journals Classification of Brain Tumor IDH Status using MRI and Deep Learning

2019 ◽  
Author(s):  
Sahil Nalawade ◽  
Gowtham Murugesan ◽  
Maryam Vejdani-Jahromi ◽  
Ryan A. Fisicaro ◽  
Chandan Ganesh Bangalore Yogananda ◽  
...  
Keyword(s):  
Deep Learning ◽  
Classification Accuracy ◽  
Test Accuracy ◽  
Low Grade ◽  
Idh Mutation ◽  
Validation Data ◽  
Mutation Status ◽  
Radiologic Imaging ◽  
Address Data

AbstractIsocitrate dehydrogenase (IDH) mutation status is an important marker in glioma diagnosis and therapy. We propose a novel automated pipeline for predicting IDH status noninvasively using deep learning and T2-weighted (T2w) MR images with minimal preprocessing (N4 bias correction and normalization to zero mean and unit variance). T2w MRI and genomic data were obtained from The Cancer Imaging Archive dataset (TCIA) for 260 subjects (120 High grade and 140 Low grade gliomas). A fully automated 2D densely connected model was trained to classify IDH mutation status on 208 subjects and tested on another held-out set of 52 subjects, using 5-fold cross validation. Data leakage was avoided by ensuring subject separation during the slice-wise randomization. Mean classification accuracy of 90.5% was achieved for each axial slice in predicting the three classes of no tumor, IDH mutated and IDH wild-type. Test accuracy of 83.8% was achieved in predicting IDH mutation status for individual subjects on the test dataset of 52 subjects. We demonstrate a deep learning method to predict IDH mutation status using T2w MRI alone. Radiologic imaging studies using deep learning methods must address data leakage (subject duplication) in the randomization process to avoid upward bias in the reported classification accuracy.

scholarly journals On the objectivity, reliability, and validity of deep learning enabled bioimage analyses

eLife ◽  
2020 ◽  
Vol 9 ◽  
Author(s):  
Dennis Segebarth ◽  
Matthias Griebel ◽  
Nikolai Stein ◽  
Cora R von Collenberg ◽  
Corinna Martin ◽  
...  
Keyword(s):  
Deep Learning ◽  
Signal To Noise Ratio ◽  
Biological Effects ◽  
Reliability And Validity ◽  
Ground Truth ◽  
Training Data ◽  
Model Organisms ◽  
Data Annotation ◽  
Bioimage Analysis ◽  
Model Training

Bioimage analysis of fluorescent labels is widely used in the life sciences. Recent advances in deep learning (DL) allow automating time-consuming manual image analysis processes based on annotated training data. However, manual annotation of fluorescent features with a low signal-to-noise ratio is somewhat subjective. Training DL models on subjective annotations may be instable or yield biased models. In turn, these models may be unable to reliably detect biological effects. An analysis pipeline integrating data annotation, ground truth estimation, and model training can mitigate this risk. To evaluate this integrated process, we compared different DL-based analysis approaches. With data from two model organisms (mice, zebrafish) and five laboratories, we show that ground truth estimation from multiple human annotators helps to establish objectivity in fluorescent feature annotations. Furthermore, ensembles of multiple models trained on the estimated ground truth establish reliability and validity. Our research provides guidelines for reproducible DL-based bioimage analyses.

Sign in / Sign up

Export Citation Format

Share Document