Search

This paper presents a novel sound event detection (SED) system for rare events occurring in an open environment. Wavelet multiresolution analysis (MRA) is used to decompose the input audio clip of 30 seconds into five levels. Wavelet denoising is then applied on the third and fifth levels of MRA to filter out the background. Significant transitions, which may represent the onset of a rare event, are then estimated in these two levels by combining the peak-finding algorithm with the K-medoids clustering algorithm. The small portions of one-second duration, called 'chunks' are cropped from the input audio signal corresponding to the estimated locations of the significant transitions. Features from these chunks are extracted by the wavelet scattering network (WSN) and are given as input to a support vector machine (SVM) classifier, which classifies them. The proposed SED framework produces an error rate comparable to the SED systems based on convolutional neural network (CNN) architecture. Also, the proposed algorithm is computationally efficient and lightweight as compared to deep learning models, as it has no learnable parameter. It requires only a single epoch of training, which is 5, 10, 200, and 600 times lesser than the models based on CNNs and deep neural networks (DNNs), CNN with long short-term memory (LSTM) network, convolutional recurrent neural network (CRNN), and CNN respectively. The proposed model neither requires concatenation with previous frames for anomaly detection nor any additional training data creation needed for other comparative deep learning models. It needs to check almost 360 times fewer chunks for the presence of rare events than the other baseline systems used for comparison in this paper. All these characteristics make the proposed system suitable for real-time applications on resource-limited devices. [ABSTRACT FROM AUTHOR]

The recognition of food images is of great significance for nutrition monitoring, food retrieval and food recommendation. However, the accuracy of recognition had not been high enough due to the complex background of food images and the characteristics of small inter-class differences and large intra-class differences. To solve these problems, this paper proposed a food image recognition method based on transfer learning and ensemble learning. Firstly, generic image features were extracted by using the convolutional neural network models (VGG19, ResNet50, MobileNet V2, AlexNet) pre-trained on the ImageNet dataset. Secondly, the 4 pre-trained models were transferred to the food image dataset for model fine-tuning. Finally, different basic learner combination strategies were adopted to establish the ensemble model and classify feature information. In this paper, several kinds of experiments were performed to compare the results of food image recognition between single models and ensemble models on food-11 dataset. The experimental results demonstrated that the accuracy of the ensemble model was the highest, reaching 96.88%, which was superior to any base learner. Therefore, the convolutional neural network model based on transfer learning and ensemble learning has strong learning ability and generalization ability, and it is feasible and practical to apply the method to food image recognition. [ABSTRACT FROM AUTHOR]

Long short-term memory (LSTM) has been effectively used to represent sequential data in recent years. However, LSTM still struggles with capturing the long-term temporal dependencies. In this paper, we propose an hourglass-shaped LSTM that is able to capture long-term temporal correlations by reducing the feature resolutions without data loss. We have used skip connections in non-adjacent layers to avoid gradient decay. In addition, an attention process is incorporated into skip connections to emphasize the essential spectral features and spectral regions. The proposed LSTM model is applied to speech enhancement and recognition applications. The proposed LSTM model uses no future information, resulting in a causal system suitable for real-time processing. The combined spectral feature sets are used to train the LSTM model for improved performance. Using the proposed model, the ideal ratio mask (IRM) is estimated as a training objective. The experimental evaluations using short-time objective intelligibility (STOI) and perceptual evaluation of speech quality (PESQ) have demonstrated that the proposed model with robust feature representation obtained higher speech intelligibility and perceptual quality. With the TIMIT, LibriSpeech, and VoiceBank datasets, the proposed model improved STOI by 16.21%, 16.41%, and 18.33% over noisy speech, whereas PESQ is improved by 31.1%, 32.9%, and 32%. In seen and unseen noisy situations, the proposed model outperformed existing deep neural networks (DNNs), including baseline LSTM, feedforward neural network (FDNN), convolutional neural network (CNN), and generative adversarial network (GAN). With the Kaldi toolkit for automated speech recognition (ASR), the proposed model significantly reduced the word error rates (WERs) and reached an average WER of 15.13% in noisy backgrounds. [ABSTRACT FROM AUTHOR]

In order to further improve the accuracy of the video-based behavior recognition method, an effective behavior recognition method in the video session using convolutional neural network is proposed. Specifically, by adding the target detection phase before the behavior recognition algorithm, the body region in the video can be accurately extracted to reduce the interference of redundant and unnecessary background noises, and at the same time, the inappropriate images can be replaced, which has reached the role of balance background trade-off, and finally, the neural network can learn the human behavior information with emphasis. By adding fragmentation and stochastic sampling, the long-time time-domain modeling of the whole video session can be established, so that the model can obtain video-level expression ability. Finally, the improved loss function is used for behavior recognition to solve the problem of classification difficulty and possible sample imbalance. In addition, we conducted the hyperparametric experiment, the ablation experiment and the contrast experiment on different open source and benchmark datasets. Compared with other commonly used behavior recognition algorithms, the experimental results verify the effectiveness of the proposed method. In addition, the related deep learning-based methods used in behavior recognition are reviewed at the beginning of this paper, and the challenges in behavior recognition and future research directions are prospected at the end of this paper, which will undoubtedly play a double role in the work of later researchers. [ABSTRACT FROM AUTHOR]

The deep neural network is used to establish a neural network model to solve the problems of low accuracy and poor accuracy of traditional algorithms in screening differentially expressed genes and function prediction during the walnut endocarp hardening stage. The paper walnut is used as the research object to analyze the biological information of paper walnut. The changes of lignin deposition during endocarp hardening from 50 days to 90 days are observed by microscope. Then, the Convolutional Neural Network (CNN) and Long and Short-term Memory (LSTM) network model are adopted to construct an expression gene screening and function prediction model. Then, the transcriptome and proteome sequencing and biological information of walnut endocarp samples at 50, 57, 78, and 90 days after flowering are analyzed and taken as the training data set of the CNN + LSTM model. The experimental results demonstrate that the endocarp of paper walnut began to harden at 57 days, and the endocarp tissue on the hardened inner side also began to stain. This indicates that the endocarp hardened laterally from outside to inside. The screening and prediction results show that the CNN + LSTM model's highest accuracy can reach 0.9264. The Accuracy, Precision, Recall, and F1-score of the CNN + LSTM model are better than the traditional machine learning algorithm. Moreover, the Receiver Operating Curve (ROC) area enclosed by the CNN + LSTM model and coordinate axis is the largest, and the Area Under Curve (AUC) value is 0.9796. The comparison of ROC and AUC proves that the CNN + LSTM model is better than the traditional algorithm for screening differentially expressed genes and function prediction in the walnut endocarp hardening stage. Using deep learning to predict expressed genes' function accurately can reduce the breeding cost and significantly improve the yield and quality of crops. This research provides scientific guidance for the scientific breeding of paper walnut. [ABSTRACT FROM AUTHOR]

Mechanical cues such as stresses and strains are now recognized as essential regulators in many biological processes like cell division, gene expression or morphogenesis. Studying the interplay between these mechanical cues and biological responses requires experimental tools to measure these cues. In the context of large scale tissues, this can be achieved by segmenting individual cells to extract their shapes and deformations which in turn inform on their mechanical environment. Historically, this has been done by segmentation methods which are well known to be time consuming and error prone. In this context however, one doesn't necessarily require a cell-level description and a coarse-grained approach can be more efficient while using tools different from segmentation. The advent of machine learning and deep neural networks has revolutionized the field of image analysis in recent years, including in biomedical research. With the democratization of these techniques, more and more researchers are trying to apply them to their own biological systems. In this paper, we tackle a problem of cell shape measurement thanks to a large annotated dataset. We develop simple Convolutional Neural Networks (CNNs) which we thoroughly optimize in terms of architecture and complexity to question construction rules usually applied. We find that increasing the complexity of the networks rapidly no longer yields improvements in performance and that the number of kernels in each convolutional layer is the most important parameter to achieve good results. In addition, we compare our step-by-step approach with transfer learning and find that our simple, optimized CNNs give better predictions, are faster in training and analysis and don't require more technical knowledge to be implemented. Overall, we offer a roadmap to develop optimized models and argue that we should limit the complexity of such models. We conclude by illustrating this strategy on a similar problem and dataset. [ABSTRACT FROM AUTHOR]

Cloud Data Computing (CDC) is conducive to precise energy-saving management of user data centers based on the real-time energy consumption monitoring of Information Technology equipment. This work aims to obtain the most suitable energy-saving strategies to achieve safe, intelligent, and visualized energy management. First, the theory of Convolutional Neural Network (CNN) is discussed. Besides, an intelligent energy-saving model based on CNN is designed to ameliorate the variable energy consumption, load, and power consumption of the CDC data center. Then, the core idea of the policy gradient (PG) algorithm is introduced. In addition, a CDC task scheduling model is designed based on the PG algorithm, aiming at the uncertainty and volatility of the CDC scheduling tasks. Finally, the performance of different neural network models in the training process is analyzed from the perspective of total energy consumption and load optimization of the CDC center. At the same time, simulation is performed on the CDC task scheduling model based on the PG algorithm to analyze the task scheduling demand. The results demonstrate that the energy consumption of the CNN algorithm in the CDC energy-saving model is better than that of the Elman algorithm and the ecoCloud algorithm. Besides, the CNN algorithm reduces the number of virtual machine migrations in the CDC energy-saving model by 9.30% compared with the Elman algorithm. The Deep Deterministic Policy Gradient (DDPG) algorithm performs the best in task scheduling of the cloud data center, and the average response time of the DDPG algorithm is 141. In contrast, the Deep Q Network algorithm performs poorly. This paper proves that Deep Reinforcement Learning (DRL) and neural networks can reduce the energy consumption of CDC and improve the completion time of CDC tasks, offering a research reference for CDC resource scheduling. [ABSTRACT FROM AUTHOR]

Siamese-based trackers have achieved excellent performance and attracted extensive attention, which regard the tracking task as a similarity learning between the target template and search regions. However, most Siamese-based trackers do not effectively exploit correlations of the spatial and channel-wise information to represent targets. Meanwhile, the cross-correlation is a linear matching method and neglects the structured and part-level information. In this paper, we propose a novel tracking algorithm for feature extraction of target templates and search region images. Based on convolutional neural networks and shuffle attention, the tracking algorithm computes the similarity between the template and a search region through a graph attention matching. The proposed tracking algorithm exploits the correlations between the spatial and channel-wise information to highlight the target region. Moreover, the graph matching can greatly alleviate the influences of appearance variations such as partial occlusions. Extensive experiments demonstrate that the proposed tracking algorithm achieves excellent tracking results on multiple challenging benchmarks. Compared with other state-of-the-art methods, the proposed tracking algorithm achieves excellent tracking performance. [ABSTRACT FROM AUTHOR]

Visual tracking is a fundamental research task in vision computer. It has broad application prospects, such as military defense and civil security. Visual tracking encounters many challenges in practical application, such as occlusion, fast motion and background clutter. Siamese based trackers achieve superior tracking performance in balanced accuracy and tracking speed. The deep feature extraction with Convolutional Neural Network (CNN) is an essential component in Siamese tracking framework. Although existing trackers take full advantage of deep feature information, the spatial structure and semantic information are not adequately exploited, which are helpful for enhancing target representations. The lack of these spatial and semantic information may lead to tracking drift. In this paper, we design a CNN feature extraction subnetwork based on a Depthwise Over-parameterized Convolutional layer (DO-Conv). A joint convolution method is introduced, namely the conventional and depthwise convolution. The depthwise convolution kernel explores independent channel information, which effectively extracts shallow spatial information and deep semantic information, and discards background information. Based on DO-Conv, we propose a novel tracking algorithm in Siamese framework (named DOSiam). Extensive experiments conducted on five benchmarks including OTB2015, VOT2016, VOT2018, GOT-10k and VOT2019-RGBT(TIR) show that the proposed DOSiam achieves leading tracking performance with real-time tracking speed at 60 FPS against state-of-the-art trackers. [ABSTRACT FROM AUTHOR]

Pneumonia remains the leading infectious cause of death in children under the age of five, killing about 700,000 children each year and affecting 7% of the world's population. X-ray images of lung become the key to the diagnosis of this disease, skilled doctors in the diagnosis of a certain degree of subjectivity, if the use of computer-aided medical diagnosis to automatically detect lung abnormalities, will improve the accuracy of diagnosis. This research aims to introduce a deep learning technology based on the combination of Xception neural network and long-term short-term memory (LSTM), which can realize automatic diagnosis of patients with pneumonia in X-ray images. First, the model uses the Xception network to extract the deep features of the data, passes the extracted features to the LSTM, and then the LSTM detects the extracted features, and finally selects the most needed features. Secondly, in the training set samples, the traditional cross-entropy loss cannot more balance the mismatch between categories. Therefore, this research combines Pearson's feature selection ideas, fusion of the correlation between the two loss functions, and optimizes the problem. The experimental results show that the accuracy rate of this paper is 96%, the receiver operator characteristic curve accuracy rate is 99%, the precision rate is 98%, the recall rate is 91%, and the F1 score accuracy rate is 94%. Compared with the existing technical methods, the research has achieved expected results on the currently available datasets. And assist doctors to provide higher reliability in the classification task of childhood pneumonia. [ABSTRACT FROM AUTHOR]

In this paper we investigate the utility of one-dimensional convolutional neural network (CNN) models in epidemiological forecasting. Deep learning models, in particular variants of recurrent neural networks (RNNs) have been studied for ILI (Influenza-Like Illness) forecasting, and have achieved a higher forecasting skill compared to conventional models such as ARIMA. In this study, we adapt two neural networks that employ one-dimensional temporal convolutional layers as a primary building block—temporal convolutional networks and simple neural attentive meta-learners—for epidemiological forecasting. We then test them with influenza data from the US collected over 2010-2019. We find that epidemiological forecasting with CNNs is feasible, and their forecasting skill is comparable to, and at times, superior to, plain RNNs. Thus CNNs and RNNs bring the power of nonlinear transformations to purely data-driven epidemiological models, a capability that heretofore has been limited to more elaborate mechanistic/compartmental disease models. [ABSTRACT FROM AUTHOR]