Showing papers in "Biomedical Signal Processing and Control in 2020"
TL;DR: A new architecture is introduced, which extracts mel-frequency cepstral coefficients, chromagram, mel-scale spectrogram, Tonnetz representation, and spectral contrast features from sound files and uses them as inputs for the one-dimensional Convolutional Neural Network for the identification of emotions using samples from the Ryerson Audio-Visual Database of Emotional Speech and Song, Berlin, and EMO-DB datasets.
502 citations
TL;DR: A deep neural network is first pre-trained as a discriminator in a generative adversarial network on different datasets of MR images to extract robust features and to learn the structure of MR pictures in its convolutional layers.
359 citations
TL;DR: This paper aims to comprehensively summarize feature descriptions and their interpretations in characterizing epileptic seizures using EEG signals, as well as to review classification performance metrics.
313 citations
TL;DR: It is believed that real-time symptom data would allow these five algorithms to provide effective and accurate identification of potential cases of COVID-19, and the framework would then document the treatment response for each patient who has contracted the virus.
303 citations
TL;DR: The aim of this paper is to conduct a detailed analysis of the different denoising techniques used for medical imaging modalities which include the 2D/3D Ultrasound (US), Magnetic Resonance (MR), Computed Tomography and Positron Emission Tomography (PET) images.
247 citations
TL;DR: An approach based on deep learning that combined convolutional neural networks (CNNs) and long short-term memory networks (LSTM) to automatically identify six types of ECG signals that had robust generalization performance and could be used as an auxiliary tool to help clinicians diagnose arrhythmia after training with a larger database.
238 citations
TL;DR: A comprehensive review for non-invasive cuff-less blood pressure estimation using the PPG approach along with their challenges and limitations is provided.
235 citations
TL;DR: A selective kernel (SK) U-Net convolutional neural network to adjust network’s receptive fields via an attention mechanism, and fuse feature maps extracted with dilated and conventional convolutions for breast mass segmentation in ultrasound (US).
230 citations
TL;DR: The results support that considering the latest two-minute raw single-channel EEG can be a reasonable choice for sleep scoring via deep neural networks with efficiency and reliability and that introducing intra-epoch temporal context learning with a deep residual network contributes to the improvement in the overall performance.
204 citations
TL;DR: This work presents a novel CNN model to extract features from retinal fundus images for better classification performance and results indicate that the proposed feature extraction technique along with the J48 classifier outperforms all the other classifiers for MESSIDOR, IDRiD, and KAGGLE datasets.
191 citations
TL;DR: The results show that it is possible to detect amusement, sadness, and neutral emotions using only galvanic skin response features and the system was able to recognize the three target emotions with accuracy up to 100% when evaluated on the test data set.
TL;DR: The experimental results demonstrated that the proposed emotion recognition system based on SRU network and ensemble learning could achieve satisfactory identification performance with relatively economic computational cost.
TL;DR: A detailed review of the key technologies related to the use of sEMG signals in human-robot interaction systems (HRISs) and the bottlenecks hindering the application are discussed.
TL;DR: An automated classification of emotions-labeled EEG signals using nonlinear higher order statistics and deep learning algorithm has the potential for accurate and rapid recognition of human emotions.
TL;DR: This work proposes a novel feature selection technique to incorporate prior information about data distribution in the RFE process as compared to the local approach of feature selection in SVM-RFE, and provides improvement over SVM -RFE in classification of control normal, mild cognitive impairment, and Alzheimer's disease subjects.
TL;DR: A network architecture named as residual cyclic unpaired encoder-decoder network (RescueNet) is designed using residual and mirroring principles to segment the whole tumor followed by core and enhance regions in a brain MRI scan.
TL;DR: A novel ternary pattern and discrete wavelet (TP-DWT) based iterative feature extraction method is proposed and a sEMG signal recognition method is presented to automate the control of prosthetic hands through surface electromyogram signals and machine learning techniques.
TL;DR: A judicious distinction between different workload levels at higher accuracy will essentially increase the performance of an operator, which effectively improves the efficiency of the Brain-Computer Interface (BCI) systems.
TL;DR: The experimental results show that the proposed multimodal image fusion scheme outperforms with some others methods by performing qualitative and quantitative analysis.
TL;DR: Simulation outcomes conducted on different types of medical images disclose that the proposed scheme demonstrates superior transparency and robustness against signal and compression attacks compared with the related hybrid optimized algorithms.
TL;DR: The database will be publicly available, and researchers can use the dataset for analyzing signals for their own proposed method in the literature and to classify EEG signals based on the arousal-valence emotion dimension and positive/negative emotions.
TL;DR: A detailed empirical exploration on Deep Neural Network (DNN) based classification system for the upper limb position invariant myoelectric signal and demonstrates that DNN based system can outperform the other existing classifiers such as k-Nearest Neighbour (kNN), Random Forest, and Decision Tree.
TL;DR: The experimental results show that the proposed improved CAD model for the classification of breast masses into the normal or abnormal and benign or malignant category is superior to other state-of-the-art models in terms of classification accuracy with a significantly reduced number of features.
TL;DR: The results of this research demonstrate that epilepsy diagnosis with quite high accuracy can be achieved with (5-12-3) MLP ANN implemented on FPGA, and show the steps towards appropriate implementation of ANN on theFPGA.
TL;DR: This study presents a CNN model that accurately identifies CHF on the basis of one raw electrocardiogram (ECG) heartbeat only, also juxtaposing existing methods typically grounded on Heart Rate Variability.
TL;DR: A novel data augmentation strategy based on duplication, concatenation and resampling of ECG episodes to balance the number of samples among different categories as well as to increase the diversity of samples is proposed.
TL;DR: An efficient method based on generative adversarial network is proposed to reduce the speckle noise and preserve the texture details in OCT images and achieves a better denoising effectiveness.
TL;DR: The single-lead ECG signal is divided into 1-min segments, and separated into frequency bands using Fourier decomposition method, which makes it computationally efficient and can be used for real-time sleep apnea detection.
TL;DR: A novel computer aided diagnostic technique based on the discrete wavelet transform (DWT) and arithmetic coding to differentiate epileptic seizure signals from normal (seizure-free) signals and has the potential for efficient application as an adjunct for the clinical diagnosis of epilepsy.
TL;DR: Compared with previous related work and clinicians, the performance of the approach can effective diagnosis CIN2+ and comparable with a senior physician, which proves the feasibility and promising of the proposed computer-aided diagnostic method.