• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.10A
• /
• pp.988-995
• /
• 2006

In this Paper, we Present a nonharmonic may geometry design method using Euclidan minimum distance function in difference Phase spaces for 2-D (azimuth/elevation) multiple baseline antenna may which has a way to reduce the number of sensor antennas while maintaining accurate DOA estimate. The major advantages of our approach is that even the shortest interelement spacing can be larger than half-wavelength and is not limit13d to linear and it can be applied successfully to any array configuration. In multiple signals impinging situation, the performance simulation results of superresolution algorithms shows the effectiveness of the proposed method. Also the 2-D asymmetric may using the Proposed method is designed and the Performance of the manufactured away through the experimental test is verified.

• Journal of Broadcast Engineering
• /
• v.2 no.2
• /
• pp.216-224
• /
• 1997

In this paper, we propose a depth information extraction method for intermediate image synthesis. As stereo matching method, MBS(Multiple-Baseline Stereo) method has been proposed, in which the matching accuracy increases by using the multiple camera, but there are some inherent problems such as computational complexity, boundary overreach(BO) in depth map, and occlusion. So, we propose the modified version of MBS so called Multi-Resolution MBS(MR-MBS). Moreover, we also propose an adaptive occlusion area processing technique to improve the accuracy of the depth information in occlusion area.

• Proceedings of the Korea Committee for Ocean Resources and Engineering Conference
• /
• 2004.11a
• /
• pp.204-209
• /
• 2004

It is well known fact that acoustic positioning systems are absolutely needed for various underwater operations. According to the distances between their sensors they are classified into three parts: long baseline(LBL), short baseline(SBL), and ultra-short baseline(USBL). Among them the USBL system is widely used because of its simplicity, although it is the most inaccurate. Recently, in order to increase the positioning accuracy, various USBL systems using broadband signal such as MFSK(Multiple Frequency Shift Keying) are produced. However, their positioning accuracy is still limited by background noise and reflected waves. Therefore, there is difficulty in applying the USBL system using MFSK signal in a shallow water with noisy conditions. In order to examine the effect of the noise and wave reflections this paper analyze position errors for various conditions using numerical simulations. The simulation results say that tile SNR must be greater than 20dB and errors in the vertical direction are slightly increased by wave reflections by upper and lower boundaries.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1997.11a
• /
• pp.161-164
• /
• 1997

본 논문에서는 임의 시점 영상 합성을 위한 깊이 정보 추출에 관한 방법을 제안한다. 깊이 정보의 추출을 위한 방법으로서 기존의 MBS(Multiple-Baseline Stereo) 방법의 깊이 맵의 경계선 연장(boundary overreach) 문제를 감소시키며 처리 시간을 개선하는 방법으로서 계층적 방법인 MR-MBS(Multi-Resolution MBS) 방법을 제시한다. 또한 MBS 방법에서 고려하지 않고 있는 폐색 영역에 대한 적절한 처리 방법을 제시한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.794-815
• /
• 2023

The rapid development of mobile communication not only has made the industry gradually diversified, but also has enhanced the service quality requirements of users. In this regard, it is imperative to consider jointly network slicing and mobile edge computing. The former mainly ensures the requirements of varied vertical services preferably, and the latter solves the conflict between the user's own energy and harsh latency. At present, the integration of the two faces many challenges and need to carry out at different levels. The main target of the paper is to minimize the energy consumption of the system, and introduce a multi-slice joint task offloading and resource allocation scheme for massive multiple input multiple output enabled heterogeneous networks. The problem is formulated by collaborative optimizing offloading ratios, user association, transmission power and resource slicing, while being limited by the dissimilar latency and rate of multi-slice. To solve it, assign the optimal problem to two sub-problems of offloading decision and resource allocation, then solve them separately by exploiting the alternative optimization technique and Karush-Kuhn-Tucker conditions. Finally, a novel slices task offloading and resource allocation algorithm is proposed to get the offloading and resource allocation strategies. Numerous simulation results manifest that the proposed scheme has certain feasibility and effectiveness, and its performance is better than the other baseline scheme.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.10
• /
• pp.11-21
• /
• 2011

This paper introduces an SIMD(Single Instruction Multiple Data) based parallel processor that efficiently processes massive data inherent in multimedia. In addition, this paper implements MMX(MultiMedia eXtension)-type instructions on the data parallel processor and evaluates and analyzes the performance of the MMX-type instructions. The reference data parallel processor consists of 16 processors each of which has a 32-bit datapath. Experimental results for a JPEG compression application with a 1280x1024 pixel image indicate that MMX-type instructions achieves a 50% performance improvement over the baseline instructions on the same data parallel architecture. In addition, MMX-type instructions achieves 100% and 51% improvements over the baseline instructions in energy efficiency and area efficiency, respectively. These results demonstrate that multimedia specific instructions including MMX-type have potentials for widely used many-core GPU(Graphics Processing Unit) and any types of parallel processors.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.408-410
• /
• 2021

In the NLP field, the pre-training model BERT launched by the Google team in 2018 has shown amazing results in various tasks in the NLP field. Subsequently, many variant models have been derived based on the original BERT, such as RoBERTa, ERNIEBERT and so on. In this paper, the WWMBERT (Whole Word Masking BERT) model suitable for Chinese text tasks was used as the baseline model of our experiment. The experiment is mainly for "Text-level Chinese text classification tasks" are improved, which mainly combines Tapt (Task-Adaptive Pretraining) and "Multi-Sample Dropout method" to improve the model, and compare the experimental results, experimental data sets and model scoring standards Both are consistent with the official WWMBERT model using Accuracy as the scoring standard. The official WWMBERT model uses the maximum and average values of multiple experimental results as the experimental scores. The development set was 97.70% (97.50%) on the "text-level Chinese text classification task". and 97.70% (97.50%) of the test set. After comparing the results of the experiments in this paper, the development set increased by 0.35% (0.5%) and the test set increased by 0.31% (0.48%). The original baseline model has been significantly improved.

• Journal of Korea Multimedia Society
• /
• v.19 no.5
• /
• pp.818-825
• /
• 2016

The existing R-peak detection research suggests improving the distortion of the signal such as baseline variations in ECG signals by using preprocessing techniques such as a bandpass filtering. However, preprocessing can introduce another distortion, as it can generate a false detection in the R-wave detection. In this paper, we propose an R-peak detection algorithm in ECG signal, based on primitive signal in order to detect reliably an R-peak in baseline variation. First, the proposed algorithm decides the primitive signal to represent the QRS complex in ECG signal, and by scaling the time axis and voltage axis, extracts multiple primitive signals. Second, the algorithm detects the candidates of the R-peak using the value of the voltage. Third, the algorithm measures the similarity between multiple primitive signals and the R-peak candidates. Finally, the algorithm detects the R-peak using the mean and the standard deviation of similarity. Throughout the experiment, we confirmed that the algorithm detected reliably a QRS group similar to multiple primitive signals. Specifically, the algorithm can achieve an R-peak detection rate greater than an average rate of 99.9%, based on eight records of MIT-BIH ADB used in this experiment.

• The Journal of the Acoustical Society of Korea
• /
• v.36 no.6
• /
• pp.425-435
• /
• 2017

In this paper, we propose optimization methods for speeding the feedforward network of deep neural networks using NEON SIMD (Single Instruction Multiple Data) parallel instructions and multi-core parallelization on the multi-core ARM processor. As the result of the optimization using SIMD parallel instructions, we present the amount of speed improvement and arithmetic precision stage by stage. Through the optimization using SIMD parallel instructions on the single core, we obtain $2.6{\times}$ speedup over the baseline implementation using C compiler. Furthermore, by parallelizing the single core implementation on the multi-core, we obtain $5.7{\times}{\sim}7.7{\times}$ speedup. The results we obtain show the possibility for applying the arithmetic-intensive deep neural network technology to applications on mobile devices.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2328-2344
• /
• 2022

2D human pose estimation still faces difficulty in low-resolution images. Most existing top-down approaches scale up the target human bonding box images to the large size and insert the scaled image into the network. Due to up-sampling, artifacts occur in the low-resolution target images, and the degraded images adversely affect the accurate estimation of the joint positions. To address this issue, we propose a multi-resolution input feature fusion network for human pose estimation. Specifically, the bounding box image of the target human is rescaled to multiple input images of various sizes, and the features extracted from the multiple images are fused in the network. Moreover, we introduce a guiding channel which induces the multi-resolution input features to alternatively affect the network according to the resolution of the target image. We conduct experiments on MS COCO dataset which is a representative dataset for 2D human pose estimation, where our method achieves superior performance compared to the strong baseline HRNet and the previous state-of-the-art methods.