CRN: End-to-end Convolutional Recurrent Network Structure Applied to Vehicle Classification

BackgroundIntracranial aneurysms (IAs) are common in the population and may cause death.ObjectiveTo develop a new fully automated detection and segmentation deep neural network based framework to assist neurologists in evaluating and contouring intracranial aneurysms from 2D+time digital subtraction angiography (DSA) sequences during diagnosis.MethodsThe network structure is based on a general U-shaped design for medical image segmentation and detection. The network includes a fully convolutional technique to detect aneurysms in high-resolution DSA frames. In addition, a bidirectional convolutional long short-term memory module is introduced at each level of the network to capture the change in contrast medium flow across the 2D DSA frames. The resulting network incorporates both spatial and temporal information from DSA sequences and can be trained end-to-end. Furthermore, deep supervision was implemented to help the network converge. The proposed network structure was trained with 2269 DSA sequences from 347 patients with IAs. After that, the system was evaluated on a blind test set with 947 DSA sequences from 146 patients.ResultsOf the 354 aneurysms, 316 (89.3%) were successfully detected, corresponding to a patient level sensitivity of 97.7% at an average false positive number of 3.77 per sequence. The system runs for less than one second per sequence with an average dice coefficient score of 0.533.ConclusionsThis deep neural network assists in successfully detecting and segmenting aneurysms from 2D DSA sequences, and can be used in clinical practice.

Download Full-text

Adieu features? End-to-end speech emotion recognition using a deep convolutional recurrent network

2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) ◽

10.1109/icassp.2016.7472669 ◽

2016 ◽

Cited By ~ 215

Author(s):

George Trigeorgis ◽

Fabien Ringeval ◽

Raymond Brueckner ◽

Erik Marchi ◽

Mihalis A. Nicolaou ◽

...

Keyword(s):

Emotion Recognition ◽

Recurrent Network ◽

Speech Emotion Recognition ◽

End To End

Download Full-text

Adaptive hop-by-hop cone vector-based forwarding protocol for underwater wireless sensor networks

International Journal of Distributed Sensor Networks ◽

10.1177/1550147720958305 ◽

2020 ◽

Vol 16 (9) ◽

pp. 155014772095830

Author(s):

Imran Ullah Khan ◽

Mazhar Islam ◽

Muhammad Ismail ◽

Abdul Baseer Qazi ◽

Sadeeq Jan ◽

...

Keyword(s):

Wireless Sensor Networks ◽

Sensor Networks ◽

Network Structure ◽

Routing Protocol ◽

Wireless Sensor ◽

Delivery Ratio ◽

Underwater Wireless Sensor Networks ◽

Base Angle ◽

End To End Delay ◽

End To End

In the recent past, a significant increase has been observed in the use of underwater wireless sensor networks for aquatic applications. However, underwater wireless sensor networks face several challenges including large propagation delays, high mobility, limited bandwidth, three-dimensional deployments, expensive manufacturing, and energy constraints. It is crucial for underwater wireless sensor networks to mitigate all these limitations primarily caused by the harsh underwater environment. To address some of the pertinent challenges, adaptive hop-by-hop cone vector-based forwarding routing protocol is proposed in this article which is based on the adaptive hop-by-hop vector-based forwarding. The novelty of adaptive hop-by-hop cone vector-based forwarding includes increasing the transmission reliability in sparse sensor regions by changing the base angle of the cone according to the network structure. The number of duplicate packets and end-to-end delay is also reduced because of the reduced base angle and a smart selection criterion for the potential forwarder node. The proposed routing protocol adaptively tunes the height and opening of the cone based on the network structure to effectively improve the performance of the network. Conclusively, this approach significantly reduces energy tax, end-to-end delay, and packet delivery ratio.

Download Full-text

End-to-end heart sound segmentation using deep convolutional recurrent network

Complex & Intelligent Systems ◽

10.1007/s40747-021-00325-w ◽

2021 ◽

Author(s):

Yao Chen ◽

Yanan Sun ◽

Jiancheng Lv ◽

Bijue Jia ◽

Xiaoming Huang

Keyword(s):

Short Term Memory ◽

Heart Sound ◽

Poor Performance ◽

Recurrent Network ◽

Sequence Recognition ◽

Operating Environments ◽

Heart Cycle ◽

Second Heart Sound ◽

End To End ◽

Heart Sound Segmentation

AbstractHeart sound segmentation (HSS) aims to detect the four stages (first sound, systole, second heart sound and diastole) from a heart cycle in a phonocardiogram (PCG), which is an essential step in automatic auscultation analysis. Traditional HSS methods need to manually extract the features before dealing with HSS tasks. These artificial features highly rely on extraction algorithms, which often result in poor performance due to the different operating environments. In addition, the high-dimension and frequency characteristics of audio also challenge the traditional methods in effectively addressing HSS tasks. This paper presents a novel end-to-end method based on convolutional long short-term memory (CLSTM), which directly uses audio recording as input to address HSS tasks. Particularly, the convolutional layers are designed to extract the meaningful features and perform the downsampling, and the LSTM layers are developed to conduct the sequence recognition. Both components collectively improve the robustness and adaptability in processing the HSS tasks. Furthermore, the proposed CLSTM algorithm is easily extended to other complex heart sound annotation tasks, as it does not need to extract the characteristics of corresponding tasks in advance. In addition, the proposed algorithm can also be regarded as a powerful feature extraction tool, which can be integrated into the existing models for HSS. Experimental results on real-world PCG datasets, through comparisons to peer competitors, demonstrate the outstanding performance of the proposed algorithm.

Download Full-text

End-to-End Multimodal 16-Day Hatching Eggs Classification

Symmetry ◽

10.3390/sym11060759 ◽

2019 ◽

Vol 11 (6) ◽

pp. 759

Author(s):

Lei Geng ◽

Zhen Peng ◽

Zhitao Xiao ◽

Jiangtao Xi

Keyword(s):

Deep Learning ◽

Blood Vessel ◽

Network Structure ◽

Short Term Memory ◽

Classification Problem ◽

Feature Maps ◽

Learning Network ◽

End To End ◽

Deep Learning Network ◽

Hatching Eggs

Sixteen-day hatching eggs are divided into fertile eggs, waste eggs, and recovered eggs. Because different categories may have the same characteristics, they are difficult to classify. Few existing algorithms can successfully solve this problem. To this end, we propose an end-to-end deep learning network structure that uses multiple forms of signals. First, we collect the photoplethysmography (PPG) signal of the hatching eggs to obtain heartbeat information and photograph hatching eggs with a camera to obtain blood vessel pictures. Second, we use two different network structures to process the two kinds of signals: Temporal convolutional networks are used to process heartbeat information, and convolutional neural networks (CNNs) are used to process blood vessel pictures. Then, we combine the two feature maps and use the long short-term memory (LSTM) network to model the context and recognize the type of hatching eggs. The system is then trained with our dataset. The experimental results demonstrate that the proposed end-to-end multimodal deep learning network structure is significantly more accurate than using a single modal network. Additionally, the method successfully solves the 16-day hatching egg classification problem.

Download Full-text

Sensitivity to Stimulus Irregularity Is Inherent in Neural Networks

Neural Computation ◽

10.1162/neco_a_01215 ◽

2019 ◽

Vol 31 (9) ◽

pp. 1789-1824

Author(s):

Teun van Gils ◽

Paul H. E. Tiesinga ◽

Bernhard Englitz ◽

Marijn B. Martens

Keyword(s):

Neural Networks ◽

Network Structure ◽

Recurrent Network ◽

Local Network ◽

Short Term ◽

Short Term Plasticity ◽

In Vivo Experiments ◽

Integrate Activity ◽

Neuron Dynamics

Behavior is controlled by complex neural networks in which neurons process thousands of inputs. However, even short spike trains evoked in a single cortical neuron were demonstrated to be sufficient to influence behavior in vivo. Specifically, irregular sequences of interspike intervals (ISIs) had a more reliable influence on behavior despite their resemblance to stochastic activity. Similarly, irregular tactile stimulation led to higher rates of behavioral responses. In this study, we identify the mechanisms enabling this sensitivity to stimulus irregularity (SSI) on the neuronal and network levels using simulated spiking neural networks. Matching in vivo experiments, we find that irregular stimulation elicits more detectable network events (bursts) than regular stimulation. Dissecting the stimuli, we identify short ISIs—occurring more frequently in irregular stimulations—as the main drivers of SSI rather than complex irregularity per se. In addition, we find that short-term plasticity modulates SSI. We subsequently eliminate the different mechanisms in turn to assess their role in generating SSI. Removing inhibitory interneurons, we find that SSI is retained, suggesting that SSI is not dependent on inhibition. Removing recurrency, we find that SSI is retained due to the ability of individual neurons to integrate activity over short timescales (“cell memory”). Removing single-neuron dynamics, we find that SSI is retained based on the short-term retention of activity within the recurrent network structure (“network memory”). Finally, using a further simplified probabilistic model, we find that local network structure is not required for SSI. Hence, SSI is identified as a general property that we hypothesize to be ubiquitous in neural networks with different structures and biophysical properties. Irregular sequences contain shorter ISIs, which are the main drivers underlying SSI. The experimentally observed SSI should thus generalize to other systems, suggesting a functional role for irregular activity in cortex.

Download Full-text