• Journal of Korea Multimedia Society
• /
• v.23 no.12
• /
• pp.1540-1551
• /
• 2020

Aiming at the problem that the existing human behavior recognition algorithm cannot fully utilize the multi-level spatio-temporal information of the network, a human behavior recognition algorithm based on a dense three-dimensional residual network is proposed. First, the proposed algorithm uses a dense block of three-dimensional residuals as the basic module of the network. The module extracts the hierarchical features of human behavior through densely connected convolutional layers; Secondly, the local feature aggregation adaptive method is used to learn the local dense features of human behavior; Then, the residual connection module is applied to promote the flow of feature information and reduced the difficulty of training; Finally, the multi-layer local feature extraction of the network is realized by cascading multiple three-dimensional residual dense blocks, and use the global feature aggregation adaptive method to learn the features of all network layers to realize human behavior recognition. A large number of experimental results on benchmark datasets KTH show that the recognition rate (top-l accuracy) of the proposed algorithm reaches 93.52%. Compared with the three-dimensional convolutional neural network (C3D) algorithm, it has improved by 3.93 percentage points. The proposed algorithm framework has good robustness and transfer learning ability, and can effectively handle a variety of video behavior recognition tasks.

• Phonetics and Speech Sciences
• /
• v.10 no.1
• /
• pp.33-38
• /
• 2018

Although deep neural network-based acoustic models have greatly improved the performance of automatic speech recognition (ASR), reverberation still degrades the performance of distant speech recognition in indoor environments. In this paper, we adopt the DenseNet, which has shown great performance results in image classification tasks, to improve the performance of reverberant speech recognition. The DenseNet enables the deep convolutional neural network (CNN) to be effectively trained by concatenating feature maps in each convolutional layer. In addition, we extend the concept of multi-resolution CNN to multi-resolution DenseNet for robust speech recognition in reverberant environments. We evaluate the performance of reverberant speech recognition on the single-channel ASR task in reverberant voice enhancement and recognition benchmark (REVERB) challenge 2014. According to the experimental results, the DenseNet-based acoustic models show better performance than do the conventional CNN-based ones, and the multi-resolution DenseNet provides additional performance improvement.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.5
• /
• pp.395-401
• /
• 2023

Sound Event Detection (SED) aims to identify not only sound category but also time interval for target sounds in an audio waveform. It is a critical technique in field of acoustic surveillance system and monitoring system. Recently, various models have introduced through Detection and Classification of Acoustic Scenes and Events (DCASE) Task 4. This paper explored how to design optimal parameters of DenseNet based model, which has led to outstanding performance in other recognition system. In experiment, DenseRNN as an SED model consists of DensNet-BC and bi-directional Gated Recurrent Units (GRU). This model is trained with Mean teacher model. With an event-based f-score, evaluation is performed depending on parameters, related to model architecture as well as model training, under the assessment protocol of DCASE task4. Experimental result shows that the performance goes up and has been saturated to near the best. Also, DenseRNN would be trained more effectively without dropout technique.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.3
• /
• pp.19-27
• /
• 2019

This paper proposes a new method using multi-layer overlapped windows to improve the performance of YOLO network which is vulnerable to detect small dense objects. In particular, the proposed method uses the YOLO Network based on the multi-layer overlapped windows to track small dense vehicles that approach from long distances. The method improves the detection performance for location and size of small vehicles. It allows crossing area of two multi-layer overlapped windows to track moving vehicles from a long distance to a short distance. And the YOLO network is optimized so that GPU computation time due to multi-layer overlapped windows should be reduced. The superiority of the proposed algorithm has been proved through various experiments using captured images from road surveillance cameras.

• Smart Structures and Systems
• /
• v.16 no.4
• /
• pp.607-621
• /
• 2015

Wireless sensor technology has been opened up numerous opportunities to advanced health and maintenance monitoring of civil infrastructure. Compare to the traditional tactics, it offers a better way of providing relevant information regarding the condition of building structure health at a lower price. Numerous domestic buildings, especially longer-span buildings have a low frequency response and challenging to measure using deployed numbers of sensors. The way the sensor nodes are connected plays an important role in providing the signals with required strengths. Out of many topologies, the dense and sparse topologies wireless sensor network were extensively used in sensor network applications for collecting health information. However, it is still unclear which topology is better for obtaining health information in terms of greatest components, node's size and degree. Theoretical and computational issues arising in the selection of the optimum topology sensor network for estimating coverage area with sensor placement in building structural monitoring are addressed. This work is an attempt to fill this gap in high-rise building structural health monitoring application. The result shows that, the sparse topology sensor network provides better performance compared with the dense topology network and would be a good choice for monitoring high-rise building structural health damage.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.7
• /
• pp.2454-2467
• /
• 2015

Increasing traffic and bandwidth requirements bring challenges to the next generation wireless networks (5G). As one of the main technology in 5G networks, Ultra-Dense Network (UDN) can be used to improve network coverage. In this paper, a radio over fiber based model is proposed to solve the load balancing problem in ultra-dense network. Stochastic differential game is introduced for the load balancing algorithm, and optimal load allocated to each access point (RAP) are formulated as Nash Equilibrium. It is proved that the optimal load can be achieved and the stochastic differential game based scheme is applicable and acceptable. Numerical results are given to prove the effectiveness of the optimal algorithm.

• Journal of Broadcast Engineering
• /
• v.24 no.4
• /
• pp.633-642
• /
• 2019

Recently, the convolution neural network (CNN) model at a single image super-resolution (SISR) have been very successful. The residual learning method can improve training stability and network performance in CNN. In this paper, we propose a SISR using recursive residual network architecture by introducing dense skip connections for learning nonlinear mapping from low-resolution input image to high-resolution target image. The proposed SISR method adopts a method of the recursive residual learning to mitigate the difficulty of the deep network training and remove unnecessary modules for easier to optimize in CNN layers because of the concise and compact recursive network via dense skip connection method. The proposed method not only alleviates the vanishing-gradient problem of a very deep network, but also get the outstanding performance with low complexity of neural network, which allows the neural network to perform training, thereby exhibiting improved performance of SISR method.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.14-25
• /
• 2021

Change detection, which finds changes in remote sensing images of the same location captured at different times, is very important because it is used in various applications. However, registration errors, building displacement errors, and shadow errors cause false positives. To solve these problems, we propose a novle deep convolutional network called CADNet (Change Attention Dense Siamese Network). CADNet uses FPN (Feature Pyramid Network) to detect multi-scale changes, applies a Change Attention Module that attends to the changes, and uses DenseNet as a feature extractor to use feature maps that contain both low-level and high-level features for change detection. CADNet performance measured from the Precision, Recall, F1 side is 98.44%, 98.47%, 98.46% for WHU datasets and 90.72%, 91.89%, 91.30% for LEVIR-CD datasets. The results of this experiment show that CADNet can offer better performance than any other traditional change detection method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.4
• /
• pp.1223-1236
• /
• 2014

With the extension of wireless technology, vehicular ad hoc networks provide important services for the dissemination of general data and emergency warnings. However, since, the vehicle topology frequently changes from a dense to a sparse network depending on the speed of the moving vehicles and the time of day, vehicular ad hoc networks require a protocol that can facilitate the efficient and reliable dissemination of emergency messages in a highly mobile environment under dense or intermittent vehicular connectivity. Therefore, this paper proposes a new vehicular broadcast protocol, called BL-CAST, that can operate effectively in both dense and sparse network scenarios. As a low overhead multi-hop broadcast protocol, BL-CAST does not rely on the periodic exchange of beacons for updating location information. Instead, the location information of a vehicle is included in a broadcast message to identify the last rebroadcasting vehicle in an intermittently connected network. Simulation results show that BL-CAST outperforms the DV-CAST protocol in terms of the end-to-end delay, message delivery ratio and network overhead.

• Journal of Korean Society for Geospatial Information Science
• /
• v.18 no.4
• /
• pp.93-99
• /
• 2010

Since the facility management for various facilities in urban area are conducted by general managers who have poor knowledge for surveying technology, it is not easy to trace the exact location of the facility in a short time with the GIS map only by themselves in the field. In order to improve it, VRS-RTK or SBAS DGPS system integrated with UMPC and PDA which is uploaded GIS field software are being used recently however lot of difficulties are still existed with the GPS positioning in urban area due to the lack of visible satellites, no reception of correction data and multipath error by the interruption of the high buildings and houses etc. Therefore, in this study, we applied with Network DGPS system which allows better reception of satellite signal and correction data even in dense housing areas with the use of GNSS receiver and CDMA mobile phone. Based on the analysis of field data, it was confirmed that standard deviations of the Network DGPS positioning are 0.3 to 0.84m with a very high positioning rate even in dense housing areas. Therefore, it was concluded that the Network DGPS system could be used widely to fast and accurate positioning for various facilities management works in dense housing areas in the future.