• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.2
• /
• pp.161-168
• /
• 2019

Recently, video equipments such as CCTV, which is spreading rapidly, is being used as a means to monitor and cope with abnormal situations in almost governments, companies, and households. However, in most cases, since recognizing the abnormal situation is carried out by the monitoring person, the immediate response is difficult and is used only for post-analysis. In this paper, we present the results of the development of video surveillance system that automatically recognizing the abnormal situations and sending such events to the smartphone immediately using the latest deep learning technology. The proposed system extracts skeletons from the human objects in real time using Openpose library and then recognizes the human behaviors automatically using deep learning technology. To this end, we reconstruct Openpose library, which developed in the Caffe framework, on Darknet framework to improve real-time processing. We also verified the performance improvement through experiments. The system to be introduced in this paper has accurate and fast behavioral recognition performance and scalability, so it is expected that it can be used for video surveillance systems for various applications.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.6 no.9
• /
• pp.59-70
• /
• 2016

The rapid development of the Chinese version of Twitter, the groom Weibo has become an important communication means for Chinese SNS users to obtain and share information. As a result, in China, the phenomenon of the power shift has emerged from the traditional opinion leaders to SNS opinion leasers. The relationship analysis of demographic variables of the Chinese SNS users and their Information on the relationship between keywords was made by utilizing the centrality analysis using Social Network Program NetMiner. China's SNS opinion leaders have general interest in daily activities with their families or friends rather than in social issues. And in case of SNS opinion leaders of high betweenness centrality, it was analyzed that general users was a key mediator role that organically out lead to the adjacent information. These properties are not independent of demographic variables, such as professional, therefore, the demographic characteristics of SNS opinion leaders showed a significant effect on the parameters of betweenness centrality. This study analyzed the characteristics of SNS users, especially opinion leaders in China by looking at the aspects of Chinese social phenomenon in terms of information. Through this study, we expect to provide basic information about the social characteristics of China through collective communication.

• Journal of KIISE:Information Networking
• /
• v.32 no.3
• /
• pp.279-290
• /
• 2005

Recently, as the serious damage caused by DDoS attacks increases, the rapid detection and the proper response mechanisms are urgent. However, existing security mechanisms do not effectively defend against these attacks, or the defense capability of some mechanisms is only limited to specific DDoS attacks. In this paper, we propose a detection architecture against DDoS attack using data mining technology that can classify the latest types of DDoS attack, and can detect the modification of existing attacks as well as the novel attacks. This architecture consists of a Misuse Detection Module modeling to classify the existing attacks, and an Anomaly Detection Module modeling to detect the novel attacks. And it utilizes the off-line generated models in order to detect the DDoS attack using the real-time traffic. We gathered the NetFlow data generated at an access router of our network in order to model the real network traffic and test it. The NetFlow provides the useful flow-based statistical information without tremendous preprocessing. Also, we mounted the well-known DDoS attack tools to gather the attack traffic. And then, our experimental results show that our approach can provide the outstanding performance against existing attacks, and provide the possibility of detection against the novel attack.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.