Violence detection is very important for public safety.However,violence detection is not an easy task.Because recognizing violence in surveillance video requires not only spatial information but also sufficient tempor...Violence detection is very important for public safety.However,violence detection is not an easy task.Because recognizing violence in surveillance video requires not only spatial information but also sufficient temporal information.In order to highlight the time information,we propose an efficient deep learning architecture for violence detection based on temporal attention mechanism,which utilizes pre-trained MobileNetV3,convolutional LSTM and temporal attention block Temporal Adaptive(TA).TA block can focus on further refining temporal information from spatial information extracted from backbone.Experimental results show the proposed model is validated on three publicly datasets:Hockey Fight,Movies,and RWF-2000 datasets.展开更多
The global growth of the Internet and the rapid expansion of social networks such as Facebook make multilingual sentiment analysis of social media content very necessary. This paper performs the first sentiment analys...The global growth of the Internet and the rapid expansion of social networks such as Facebook make multilingual sentiment analysis of social media content very necessary. This paper performs the first sentiment analysis on code-mixed Bambara-French Facebook comments. We develop four Long Short-term Memory（LSTM）-based models and two Convolutional Neural Network（CNN）-based models, and use these six models, Na？ve Bayes, and Support Vector Machines（SVM） to conduct experiments on a constituted dataset. Social media text written in Bambara is scarce. To mitigate this weakness, this paper uses dictionaries of character and word indexes to produce character and word embedding in place of pre-trained word vectors. We investigate the effect of comment length on the models and perform a comparison among them. The best performing model is a one-layer CNN deep learning model with an accuracy of 83.23 %.展开更多
Recently,deep learning(DL)has been widely used in the field of remaining useful life(RUL)prediction.Among various DL technologies,recurrent neural network(RNN)and its variant,e.g.,long short-term memory(LSTM)network,h...Recently,deep learning(DL)has been widely used in the field of remaining useful life(RUL)prediction.Among various DL technologies,recurrent neural network(RNN)and its variant,e.g.,long short-term memory(LSTM)network,have gained extensive attention for their ability to capture temporal dependence.Although existing RNN-based methods have demonstrated their RUL prediction effectiveness,they still suffer from the following two limitations:1)it is difficult for the RNN to directly extract degradation features from original monitoring data and 2)most RNN-based prognostics methods are unable to quantify RUL uncertainty.To address the aforementioned limitations,this paper proposes a new prognostics method named residual convolution LSTM(RC-LSTM)network.In the RC-LSTM,a new ResNet-based convolution LSTM(Res-ConvLSTM)layer is stacked with a convolution LSTM(ConvLSTM)layer to extract degradation representations from monitoring data.Then,under the assumption that the RUL follows a normal distribution,an appropriate output layer is constructed to quantify the uncertainty of prediction results.Finally,the effectiveness and superiority of the RC-LSTM are verified using monitoring data from accelerated bearing degradation tests.展开更多
Trajectory prediction is a fundamental and challenging task for numerous applications,such as autonomous driving and intelligent robots.Current works typically treat pedestrian trajectories as a series of 2D point coo...Trajectory prediction is a fundamental and challenging task for numerous applications,such as autonomous driving and intelligent robots.Current works typically treat pedestrian trajectories as a series of 2D point coordinates.However,in real scenarios,the trajectory often exhibits randomness,and has its own probability distribution.Inspired by this observation and other movement characteristics of pedestrians,we propose a simple and intuitive movement description called a trajectory distribution,which maps the coordinates of the pedestrian trajectory to a 2D Gaussian distribution in space.Based on this novel description,we develop a new trajectory prediction method,which we call the social probability method.The method combines trajectory distributions and powerful convolutional recurrent neural networks.Both the input and output of our method are trajectory distributions,which provide the recurrent neural network with sufficient spatial and random information about moving pedestrians.Furthermore,the social probability method extracts spatio-temporal features directly from the new movement description to generate robust and accurate predictions.Experiments on public benchmark datasets show the effectiveness of the proposed method.展开更多
文摘Violence detection is very important for public safety.However,violence detection is not an easy task.Because recognizing violence in surveillance video requires not only spatial information but also sufficient temporal information.In order to highlight the time information,we propose an efficient deep learning architecture for violence detection based on temporal attention mechanism,which utilizes pre-trained MobileNetV3,convolutional LSTM and temporal attention block Temporal Adaptive(TA).TA block can focus on further refining temporal information from spatial information extracted from backbone.Experimental results show the proposed model is validated on three publicly datasets:Hockey Fight,Movies,and RWF-2000 datasets.
基金Supported by the National Natural Science Foundation of China(61272451,61572380,61772383 and 61702379)the Major State Basic Research Development Program of China(2014CB340600)
文摘The global growth of the Internet and the rapid expansion of social networks such as Facebook make multilingual sentiment analysis of social media content very necessary. This paper performs the first sentiment analysis on code-mixed Bambara-French Facebook comments. We develop four Long Short-term Memory（LSTM）-based models and two Convolutional Neural Network（CNN）-based models, and use these six models, Na？ve Bayes, and Support Vector Machines（SVM） to conduct experiments on a constituted dataset. Social media text written in Bambara is scarce. To mitigate this weakness, this paper uses dictionaries of character and word indexes to produce character and word embedding in place of pre-trained word vectors. We investigate the effect of comment length on the models and perform a comparison among them. The best performing model is a one-layer CNN deep learning model with an accuracy of 83.23 %.
基金This research was supported by National Natural Science Foundation of China(52005387,52025056)Project funded by China Postdoctoral Science Foundation(2020M673380)Fundamental Research Funds for the Central Universities.
文摘Recently,deep learning(DL)has been widely used in the field of remaining useful life(RUL)prediction.Among various DL technologies,recurrent neural network(RNN)and its variant,e.g.,long short-term memory(LSTM)network,have gained extensive attention for their ability to capture temporal dependence.Although existing RNN-based methods have demonstrated their RUL prediction effectiveness,they still suffer from the following two limitations:1)it is difficult for the RNN to directly extract degradation features from original monitoring data and 2)most RNN-based prognostics methods are unable to quantify RUL uncertainty.To address the aforementioned limitations,this paper proposes a new prognostics method named residual convolution LSTM(RC-LSTM)network.In the RC-LSTM,a new ResNet-based convolution LSTM(Res-ConvLSTM)layer is stacked with a convolution LSTM(ConvLSTM)layer to extract degradation representations from monitoring data.Then,under the assumption that the RUL follows a normal distribution,an appropriate output layer is constructed to quantify the uncertainty of prediction results.Finally,the effectiveness and superiority of the RC-LSTM are verified using monitoring data from accelerated bearing degradation tests.
基金supported in part by the National Natural Science Foundation of China under Grant Nos.61772474,61802351,61822701,and 61872324in part by the Program for Science and Technology Innovation Talents in Universities of Henan Province under Grant No.20HASTIT021.
文摘Trajectory prediction is a fundamental and challenging task for numerous applications,such as autonomous driving and intelligent robots.Current works typically treat pedestrian trajectories as a series of 2D point coordinates.However,in real scenarios,the trajectory often exhibits randomness,and has its own probability distribution.Inspired by this observation and other movement characteristics of pedestrians,we propose a simple and intuitive movement description called a trajectory distribution,which maps the coordinates of the pedestrian trajectory to a 2D Gaussian distribution in space.Based on this novel description,we develop a new trajectory prediction method,which we call the social probability method.The method combines trajectory distributions and powerful convolutional recurrent neural networks.Both the input and output of our method are trajectory distributions,which provide the recurrent neural network with sufficient spatial and random information about moving pedestrians.Furthermore,the social probability method extracts spatio-temporal features directly from the new movement description to generate robust and accurate predictions.Experiments on public benchmark datasets show the effectiveness of the proposed method.
