• Journal of Internet Computing and Services
• /
• v.21 no.4
• /
• pp.35-41
• /
• 2020

This paper proposes an algorithm to improve the compression performance of the adaptive multi-rate (AMR) speech coding using the context tree weighting (CTW) method. AMR is the voice encoding standard adopted by IMT-2000, and supports 8 transmission rates from 4.75 kbit/s to 12.2 kbit/s to cope with changes in the channel condition. CTW as a kind of the arithmetic coding, uses a variable-order Markov model. Considering that CTW operates bit by bit, we propose an algorithm that re-orders AMR data and compresses them with CTW. To verify the validity of the proposed algorithm, an experiment is conducted to compare the proposed algorithm with existing compression methods including ZIP in terms of compression ratio. Experimental results indicate that the average additional compression rate in AMR data is about 3.21% with ZIP and about 9.10% with the proposed algorithm. Thus our algorithm improves the compression performance of AMR data by about 5.89%.

• The Journal of the Korea Contents Association
• /
• v.20 no.6
• /
• pp.556-578
• /
• 2020

There are many different researches about how to analyze issues based on real-time news streams. But, there are few researches which analyze issues hierarchically from news articles and even a previous research of hierarchical issue analysis make clustering speed slower as the increment of news articles. In this paper, we propose a hierarchical and incremental clustering for semi real-time issue analysis on news articles. We trained siamese neural network based weighted cosine similarity model, applied this model to k-means algorithm which is used to make word clusters and converted news articles to document vectors by using these word clusters. Finally, we initialized an issue cluster tree from document vectors, updated this tree whenever news articles happen, and analyzed issues in semi real-time. Through the experiment and evaluation, we showed that up to about 0.26 performance has been improved in terms of NMI. Also, in terms of speed of incremental clustering, we also showed about 10 times faster than before.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.1
• /
• pp.24-31
• /
• 2020

In this paper, the deep convolutional neural network for sound event detection is studied. Especially, the end-to-end neural network, which generates the detection results from the input audio waveform, is studied for weakly supervised problem that includes weakly-labeled and unlabeled dataset. The proposed system is based on the network structure that consists of deeply-stacked 1-dimensional convolutional neural networks, and enhanced by the skip connection and gating mechanism. Additionally, the proposed system is enhanced by the sound event detection and post processings, and the training step using the mean-teacher model is added to deal with the weakly supervised data. The proposed system was evaluated by the Detection and Classification of Acoustic Scenes and Events (DCASE) 2019 Task 4 dataset, and the result shows that the proposed system has F₁-scores of 54 % (segment-based) and 32 % (event-based).

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.1
• /
• pp.8-15
• /
• 2020

In this paper, we propose a frame synchronization algorithm for robust to the combined effects of large Doppler fluctuations and extended, time-varying multipath in the underwater acoustic communication. From the algorithm, we can recover a high timing error which is occurred from an acoustic propagation delay and uncertainty of oscillator between transmitter and receiver. In order to verify the performance of the synchronization algorithm, the lake trial results are used. The lake experiments are performed in a Gyeongcheonho located in Mungyeong-si, Gyeongsangbuk-do. We can see that the start position of frame is adjusted after the frame synchronization while the receiver moving.

• Journal of Korea Technology Innovation Society
• /
• v.17 no.4
• /
• pp.703-732
• /
• 2014

Meta-evaluation for national R&D programs should be established within the framework of whole R&D system linked with science and technology needs and objectives. The R&D program evaluation system can be understood as a circulation of input, implementation, output, and utilization. In the result of the Delphi survey, which consisted of twenty-four experts (performed in three rounds), twenty-four indices were developed. The meta-evaluation model included sub-items which were important points of view of it as follow (eight items): propriety of planning, sufficiency of information, and propriety of evaluators in evaluation input; appropriateness of method and appropriateness of procedure in evaluation implementation; credibility of output in evaluation output; and usefulness of report and application of evaluation utilization. As the application of the meta-evaluation model, the total mean was lower than 3.0 of the median value of 2.73 (5 point Likert scale). Finally, it was suggested that the feedback of evaluation results should be more enforced to R&D program improvement particularly (evaluation utilization 2.50).

• Journal of Broadcast Engineering
• /
• v.20 no.2
• /
• pp.248-256
• /
• 2015

In this paper, we model the interference plus noise signal for terrestrial cloud transmission systems and present bit error rate (BER) performances. Since terrestrial cloud transmission systems experience co-channel interference from one or more transmitters, they have to operate under a negative signal-to-interference plus noise ratio (SINR) region. The interference plus noise signal can be modeled as Gaussian random variable under the required SINR region and we observe the BER performance of the cloud transmission system using the derived model. Also, we propose an improved channel estimation scheme by averaging the channel estimates based on least square based interpolation scheme. Simulation results show that the cloud transmission system can operate under negative SINR region using the proposed channel estimation scheme.

• Journal of Satellite, Information and Communications
• /
• v.9 no.3
• /
• pp.9-14
• /
• 2014

Recently, precision guided munition systems and missile defense systems based on GPS have been taking a key role in modern warfare. In warfare however, unexpected interferences cause by large/small scale fading, radio frequency interferences, etc. These interferences result in a severe GPS positioning error, which could occur late supports and friendly fires. To solve the problems, this paper proposes an interference mitigation positioning method by adopting a wavelet denoising filter algorithm. The algorithm is applied to a GPS/QZSS/Wi-Fi combined positioning system which was performed by this laboratory. Experimental results of this paper are based on a real field test data of a GPS/QZSS/Wi-Fi combined positioning system and a simulation data of a wavelet denoising filter algorithm. At the end, the simulation result demonstrates its superiority by showing a 21.6% improved result in comparison to a conventional GPS system.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.4
• /
• pp.377-385
• /
• 2015

This paper proposes a novel forward vehicle tracking algorithm based on the WMIL(Weighted Multiple Instance Learning) equipped with a particle filter. In the proposed algorithm Haar-like features are used to train a vehicle object detector to be tracked and the location of the object are obtained from the recognition result. In order to combine both the WMIL to construct the vehicle detector and the particle filter, the proposed algorithm updates the object location by executing the propagation, observation, estimation, and selection processes involved in particle filter instead of finding the credence map in the search area for every frame. The proposed algorithm inevitably increases the computation time because of the particle filter, but the tracking accuracy was highly improved compared to Ababoost, MIL(Multiple Instance Learning) and MIL-based ones so that the position error was 4.5 pixels in average for the videos of national high-way, express high-way, tunnel and urban paved road scene.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.9
• /
• pp.377-384
• /
• 2015

In this paper we propose a visualization tool for distortion-free time-series matching. Supporting distortion-free is a very important factor in time-series matching to get more accurate matching results. In this paper, we visualize the result of time-series matching, which removes various time-series distortions such as noise, offset translation, amplitude scaling, and linear trend by using moving average, normalization, linear detrending transformations, respectively. The proposed visualization tool works as a client-server model. The client sends a user-selected time-series, of which distortions are removed, to the server and visualizes the matching results. The server efficiently performs the distortion-free time-series matching on the multi-dimensional R*-tree index. By visualizing the matching result as five different charts, we can more easily and more intuitively understand the matching result.

• Journal of the Korea Convergence Society
• /
• v.11 no.4
• /
• pp.11-16
• /
• 2020

As a plan for oil spill accidents, research to collect and analyze optimal equipment assignments is essential. However, studies that have diversified and analyzed the optimal equipment assignments for responding to oil spill accidents have not been preceded. In response to the need for analyzing optimal equipment assignments study, we devised a genetic algorithm for optimal equipment assignments. The designed genetic algorithm yielded 10,000 optimal equipment assignments. We clustered using the k-means algorithm. As a result, the two clusters of Yeosu, Daesan, and Ulsan, which are expected to be the largest spills, were clearly identified. We also projected 16-dimensional data in two dimensions via Sammon's mapping. The projected data were analyzed for distribution. We confirmed that results of the simulation were better than those of optimal equipment assignments included in the cluster.In the future, it will be possible to implement an approximate model with excellent performance based on this study.