• Information and Communications Magazine
• /
• v.33 no.10
• /
• pp.65-70
• /
• 2016

최근 스마트폰의 이용 사례가 증가함에 따라, 스마트폰에 내장되어 있는 다양한 센서를 이용하여 인간의 행동을 인식하기 위한 연구가 많은 각광을 받고 있다. 본고에서는 인간의 기본적인 행동 중에 앉기, 걷기, 달리기 등의 행동 특성을 스마트폰에 내장되어 있는 3축 가속도 센서를 통하여 분석하고 인간의 기본적 행동을 자동으로 인식하기 위한 방법에 대하여 비교 분석하는 것을 목적으로 한다. 구체적으로는 스마트폰에 내장되어 있는 3차원 가속도 센서로부터 추출된 데이터를 시간축에서 샘플링하여 인간의 행동을 인식하기 위한 기댓값 최대화 알고리즘, 랜덤 포레스트, 딥러닝 기반의 기계학습 방법을 비교하여 각 기계학습 알고리즘의 장단점을 분석한다.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.4
• /
• pp.7-12
• /
• 2019

Ambient backscatter communication is limited to channel estimation technique through a pilot signal, which is a channel estimation method in current RF communication, due to transmission power efficiency. In a limited transmission power environment, the research of traditional ambient backscatter communication has been studied assuming that it is an ideal channel without signal distortions due to channel conditions. In this paper, we propose an expectation-maximization(EM) algorithm, one of the blind channel estimation techniques, as a channel estimation method in ambient backscatter communication system which is the state of channel following normal distribution. In the proposed system model, the simulations confirm that channel estimate through EM algorithm is approaching the lower bound of the mean square error compared with the Bayesian Cramer-Rao Boundary(BCRB) to check performance. It shows that the channel parameter can be estimated in the ambient backscatter communication system.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.2
• /
• pp.81-86
• /
• 2020

Ambient backscatter communication has a drawback in which the transmission power is limited because the data is transmitted using the ambient RF signal. In order to improve transmission efficiency between transceiver, a channel estimator capable of estimating channel state at a receiver is needed. In this paper, we consider the K-means algorithm to improve the performance of the channel estimator based on EM algorithm. The simulation uses MSE as a performance parameter to verify the performance of the proposed channel estimator. The initial value setting through K-means shows improved performance compared to the channel estimation method using the general EM algorithm.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.11
• /
• pp.537-542
• /
• 2017

Aggressive driving is a major cause of car accidents. Previous studies have mainly analyzed young driver's aggressive driving tendency, yet they were only done through pure clustering or classification technique of machine learning. However, since elderly people have different driving habits due to their fragile physical conditions, it is necessary to develop a new method such as enhancing the characteristics of driving data to properly analyze aggressive driving of elderly drivers. In this study, acceleration data collected from a smartphone of a driving vehicle is analyzed by a newly proposed ECA(Enhanced Clustering method for Acceleration data) technique, coupled with a conventional clustering technique (K-means Clustering, Expectation-maximization algorithm). ECA selects high-intensity data among the data of the cluster group detected through K-means and EM in all of the subjects' data and models the characteristic data through the scaled value. Using this method, the aggressive driving data of all youth and elderly experiment participants were collected, unlike the pure clustering method. We further found that the K-means clustering has higher detection efficiency than EM method. Also, the results of K-means clustering demonstrate that a young driver has a driving strength 1.29 times higher than that of an elderly driver. In conclusion, the proposed method of our research is able to detect aggressive driving maneuvers from data of the elderly having low operating intensity. The proposed method is able to construct a customized safe driving system for the elderly driver. In the future, it will be possible to detect abnormal driving conditions and to use the collected data for early warning to drivers.

• Nuclear Medicine and Molecular Imaging
• /
• v.43 no.5
• /
• pp.459-467
• /
• 2009

Purpose: The maximum likelihood-expectation maximization (ML-EM) is the statistical reconstruction algorithm derived from probabilistic model of the emission and detection processes. Although the ML-EM has many advantages in accuracy and utility, the use of the ML-EM is limited due to the computational burden of iterating processing on a CPU (central processing unit). In this study, we developed a parallel computing technique on GPU (graphic processing unit) for ML-EM algorithm. Materials and Methods: Using Geforce 9800 GTX+ graphic card and CUDA (compute unified device architecture) the projection and backprojection in ML-EM algorithm were parallelized by NVIDIA's technology. The time delay on computations for projection, errors between measured and estimated data and backprojection in an iteration were measured. Total time included the latency in data transmission between RAM and GPU memory. Results: The total computation time of the CPU- and GPU-based ML-EM with 32 iterations were 3.83 and 0.26 see, respectively. In this case, the computing speed was improved about 15 times on GPU. When the number of iterations increased into 1024, the CPU- and GPU-based computing took totally 18 min and 8 see, respectively. The improvement was about 135 times and was caused by delay on CPU-based computing after certain iterations. On the other hand, the GPU-based computation provided very small variation on time delay per iteration due to use of shared memory. Conclusion: The GPU-based parallel computation for ML-EM improved significantly the computing speed and stability. The developed GPU-based ML-EM algorithm could be easily modified for some other imaging geometries.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.294-297
• /
• 2015

Recently, big data analysis which is based on machine learning has been gained a lot of attentions in various fields. Especially, agriculture is considered as one promising field that machine learning algorithm can be efficiently utilized and accordingly, lots of works have been done so far. However, most of the researches are focusing on the forecast of weather or analysis of genome, and machine learning algorithm for livestock management, especially which uses individual data of livestocks, e.g., temperature and movement, are not properly investigated yet. In this work, we propose fast abnormal livestock detection algorithm based on machine learning, more specifically expectation maximization, such that livestock which has problem can be efficiently and promptly found. In our proposed scheme, livestocks are divided into two clusters using expectation maximization based on their bionic data and the abnormal livestock can be detected by comparing the size of two clusters. Especially, we divide the case in which single livestock has problem and the case in which livestocks have epidemic such that fast response is enabled when epidemic case. Moreover, our algorithm does not need statistical information.

• Nuclear Medicine and Molecular Imaging
• /
• v.43 no.5
• /
• pp.443-450
• /
• 2009

Purpose: Conventional image reconstruction uses simplified physical models of projection. However, real physics, for example 3D reconstruction, takes too long time to process all the data in clinic and is unable in a common reconstruction machine because of the large memory for complex physical models. We suggest the realistic distributed memory model of fast-reconstruction using parallel processing on personal computers to enable large-scale technologies. Materials and Methods: The preliminary tests for the possibility on virtual manchines and various performance test on commercial super computer, Tachyon were performed. Expectation maximization algorithm with common 2D projection and realistic 3D line of response were tested. Since the process time was getting slower (max 6 times) after a certain iteration, optimization for compiler was performed to maximize the efficiency of parallelization. Results: Parallel processing of a program on multiple computers was available on Linux with MPICH and NFS. We verified that differences between parallel processed image and single processed image at the same iterations were under the significant digits of floating point number, about 6 bit. Double processors showed good efficiency (1.96 times) of parallel computing. Delay phenomenon was solved by vectorization method using SSE. Conclusion: Through the study, realistic parallel computing system in clinic was established to be able to reconstruct by plenty of memory using the realistic physical models which was impossible to simplify.

• The Korean Journal of Nuclear Medicine
• /
• v.39 no.6
• /
• pp.464-472
• /
• 2005

Purpose: SPECT using a fan-beam collimator improves spatial resolution and sensitivity. For the reconstruction from fan-beam projections, it is necessary to implement direct fan-beam reconstruction methods without transforming the data into the parallel geometry. In this study, various fan-beam reconstruction algorithms were implemented and their performances were compared. Materials and Methods: The projector for fan-beam SPECT was implemented using a ray-tracing method. The direct reconstruction algorithms implemented for fan-beam projection data were FBP (filtered backprojection), EM (expectation maximization), OS-EM (ordered subsets EM) and MAP-EM OSL (maximum a posteriori EM using the one-step late method) with membrane and thin-plate models as priors. For comparison, the fan-beam protection data were also rebinned into the parallel data using various interpolation methods, such as the nearest neighbor, bilinear and bicubic interpolations, and reconstructed using the conventional EM algorithm for parallel data. Noiseless and noisy projection data from the digital Hoffman brain and Shepp/Logan phantoms were reconstructed using the above algorithms. The reconstructed images were compared in terms of a percent error metric. Results: for the fan-beam data with Poisson noise, the MAP-EM OSL algorithm with the thin-plate prior showed the best result in both percent error and stability. Bilinear interpolation was the most effective method for rebinning from the fan-beam to parallel geometry when the accuracy and computation load were considered. Direct fan-beam EM reconstructions were more accurate than the standard EM reconstructions obtained from rebinned parallel data. Conclusion: Direct fan-beam reconstruction algorithms were implemented, which provided significantly improved reconstructions.

• Nuclear Medicine and Molecular Imaging
• /
• v.41 no.3
• /
• pp.234-240
• /
• 2007

Purpose: In this study we propose a block-iterative method for reconstructing Compton scattered data. This study shows that the well-known expectation maximization (EM) approach along with its accelerated version based on the ordered subsets principle can be applied to the problem of image reconstruction for Compton camera. This study also compares several methods of constructing subsets for optimal performance of our algorithms. Materials and Methods: Three reconstruction algorithms were implemented; simple backprojection (SBP), EM, and ordered subset EM (OSEM). For OSEM, the projection data were grouped into subsets in a predefined order. Three different schemes for choosing nonoverlapping subsets were considered; scatter angle-based subsets, detector position-based subsets, and both scatter angle- and detector position-based subsets. EM and OSEM with 16 subsets were performed with 64 and 4 iterations, respectively. The performance of each algorithm was evaluated in terms of computation time and normalized mean-squared error. Results: Both EM and OSEM clearly outperformed SBP in all aspects of accuracy. The OSEM with 16 subsets and 4 iterations, which is equivalent to the standard EM with 64 iterations, was approximately 14 times faster in computation time than the standard EM. In OSEM, all of the three schemes for choosing subsets yielded similar results in computation time as well as normalized mean-squared error. Conclusion: Our results show that the OSEM algorithm, which have proven useful in emission tomography, can also be applied to the problem of image reconstruction for Compton camera. With properly chosen subset construction methods and moderate numbers of subsets, our OSEM algorithm significantly improves the computational efficiency while keeping the original quality of the standard EM reconstruction. The OSEM algorithm with scatter angle- and detector position-based subsets is most available.