• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권12호
• /
• pp.4476-4491
• /
• 2021

Video service providers tend to face user network problems in the process of transmitting video streams. They strive to provide user with superior video quality in a limited bitrate environment. It is necessary to accurately determine the target bitrate range of the video under different quality requirements. Recently, several schemes have been proposed to meet this requirement. However, they do not take the impact of visual influence into account. In this paper, we propose a new multi-category model to accurately predict the target bitrate range with target visual quality by machine learning. Firstly, a dataset is constructed to generate multi-category models by machine learning. The quality score ladders and the corresponding bitrate-interval categories are defined in the dataset. Secondly, several types of spatial-temporal features related to VMAF evaluation metrics and visual factors are extracted and processed statistically for classification. Finally, bitrate prediction models trained on the dataset by RandomForest classifier can be used to accurately predict the target bitrate of the input videos with target video quality. The classification prediction accuracy of the model reaches 0.705 and the encoded video which is compressed by the bitrate predicted by the model can achieve the target perceptual quality.

• Geomechanics and Engineering
• /
• 제15권3호
• /
• pp.879-888
• /
• 2018

Probability-based design codes have been developed to sufficiently confirm the safety level of structures. One of the most acceptable probability-based approaches is Load Resistance Factor Design (LRFD), which measures the safety level of the structures in terms of the reliability index. The main contribution of this paper is to calibrate the load and resistance factors of the design code for tunnels. The load and resistance factors are calculated using the available statistical models and probability-based procedures. The major steps include selection of representative structures, consideration of the limit state functions, calculation of reliability for the selected structures, selection of the target reliability index and calculation of load factors and resistance factors. The load and resistance models are reviewed. Statistical models of resistance (load carrying capacity) are summarized for strength limit state in bending, shear and compression. The reliability indices are calculated for several segments of a selected circular tunnel designed according to the tunnel manual report (Tunnel Manual). The novelty of this paper is the selection of the target reliability. In doing so, the uniform spectrum of reliability indices is proposed based on the probability paper. The final recommendation is proposed based on the closeness to the target reliability index.

• 대한조선학회논문집
• /
• 제44권6호
• /
• pp.572-578
• /
• 2007

A Rigid Inflatable Boat (RIB) is now widely used for commercial and military purpose. In this paper, it is supposed that seven-meter-class RIB be used as an unmanned target ship for naval training. In order to develop many tactical maneuvering patterns of a target ship, a simple horizontal maneuvering model of a RIB is needed. Therefore, models of speed and yaw rate are constructed as the first-order differential equations based on Lewandowski#s empirical formula for steady turning circle diameter of a conventional planning hull. Some parameters in the models are determined using the results of sea trial tests. Finally, proposed models are validated through the comparison of the simulation result with the sea trial result for a specific scenario. Even though a simple model does not represent the horizontal motion of a RIB precisely, however, it can be used enough to develop tactical trajectory patterns.

• 한국산업정보학회논문지
• /
• 제25권5호
• /
• pp.1-11
• /
• 2020

가상착용기술은 온라인 의류 쇼핑 활성화를 위해 중요한 기술이다. 최근 이미지 기반 가상착용기술은 의상과 착용 대상 신체의 3차원 정보가 필요하지 않다는 실용성 때문에 큰 관심을 받고 있다. 그러나 기존의 이미지 기반 알고리즘의 2차원 기하변형 방식의 한계로 인하여 대상 인물의 포즈와 의상 이미지의 형태가 큰 차이가 있는 경우 자연스러운 의상변형을 하지 못한다. 본 논문에서는 이러한 문제를 해결하기 위해 3차원 인체 모델을 이용하여 2차원 의상 사진으로 부터 의상의 3차원 모델을 생성하고, 대상 인물의 자세와 체형에 맞게 3차원 변형 후 렌더링하고 대상 인간 이미지와 혼합을 통하여 가상착용 이미지를 생성할 수 있다. 기존 연구에서 사용된 VITON 데이터 세트를 사용한 실험 결과는 3차원 변형이 요구되는 경우에 2차원 이미지 기반 가상착용 결과들에 비교했을 때 자연스러운 결과를 보인다.

• Nuclear Engineering and Technology
• /
• 제54권12호
• /
• pp.4698-4707
• /
• 2022

As part of the ICRP Task Group 103 project, we developed ten thyroid models for the pediatric mesh-type reference computational phantoms (MRCPs). The thyroid is not only a radiosensitive target organ needed for effective dose calculation but an important source region particularly for radioactive iodines. The thyroid models for the pediatric MRCPs were constructed by converting those of the pediatric voxel-type reference computational phantoms (VRCPs) in ICRP Publication 143 to a high-quality mesh format, faithfully maintaining their original topology. At the same time, we improved several anatomical parameters of the thyroid models for the pediatric MRCPs, including the mass, overlying tissue thickness, location, and isthmus dimensions. Absorbed doses to the thyroid for the pediatric MRCPs for photon external exposures were calculated and compared with those of the pediatric VRCPs, finding that the differences between the MRCPs and VRCPs were not significant except for very low energies (<0.03 MeV). Specific absorbed fractions (target ⟵ thyroid) for photon internal exposures were also compared, where significant differences were frequently observed especially for the target organs/tissues close to the thyroid (e.g., a factor of ~1.2-~327 for the thymus as a target) due mainly to anatomical improvement of the MRCP thyroid models.

• 품질경영학회지
• /
• 제29권3호
• /
• pp.66-78
• /
• 2001

This paper deals with problem of determining process target value under automated visual inspection(AVI) system. Three independent error sources - digitizing error, illumination error, and positional error - which have a close relationship with the performance of the AVI system, are considered. Assuming that digitizing error is uniformly or normally distributed and illumination and positional errors are normally distributed, respectively, the distribution function for the error of measured lengths is derived when the length of a product is measured by the AVI system. Then, Optimal target values under two error models of AVI system are obtained by minimizing the total expected cost function which consists of give away, rework and penalty cost. To validate two process setting models, AVI system for drinks filling process is made up and test results are discussed.

• International Journal of Control, Automation, and Systems
• /
• 제5권3호
• /
• pp.307-316
• /
• 2007

In real system application, the interacting multiple model (IMM) based algorithm operates with the following problems: it requires less computing resources as well as a good performance with respect to the various target maneuvering, it requires a robust performance with respect to the time-varying measurement noise, and further, it requires an easy design procedure in terms of its structures and parameters. To solve these problems, an adaptive fuzzy interacting multiple model (AFIMM) algorithm, which is based on the basis sub-models defined by considering the maneuvering property and the time-varying mode transition probabilities designed by using the mode probabilities as the inputs of the fuzzy decision maker whose widths are adjusted, is proposed. To verify the performance of the proposed algorithm, a radar target tracking is performed. Simulation results show that the proposed AFIMM algorithm solves all problems in the real system application of the IMM based algorithm.

• 대한원격탐사학회지
• /
• 제38권2호
• /
• pp.199-213
• /
• 2022

비지도 도메인 적응은 연단위 작물 분류를 위해 매년 반복적으로 양질의 훈련자료를 수집해야 하는 비실용적인 문제를 해결할 수 있다. 이 연구에서는 작물 분류를 위한 딥러닝 기반 비지도 도메인 적응 모델의 적용성을 평가하였다. 우리나라 마늘, 양파 주산지인 합천군과 창녕군을 대상으로 무인기 영상을 이용한 작물 분류 실험을 통해 deep adaptation network (DAN), deep reconstruction-classification network, domain adversarial neural network (DANN)의 3개의 비지도 도메인 적응 모델을 정량적으로 비교하였다. 비지도 도메인 적응 모델의 분류 성능을 평가하기 위해 소스 베이스라인 및 대상 베이스라인 모델로 convolutional neural networks (CNNs)을 추가로 적용하였다. 3개의 비지도 도메인 적응 모델은 소스 베이스라인 CNN보다 우수한 성능을 보였으나, 소스 도메인 영상과 대상 도메인 영상의 자료 분포 간 불일치 정도에 따라 서로 다른 분류 성능을 보였다. DAN의 분류 성능은 두 도메인 영상 간 불일치가 작을 때 다른 두 모델에 비해 분류 성능이 높은 반면에 DANN은 두 도메인 영상 간 불일치가 클 때 가장 우수한 분류 성능을 보였다. 따라서 신뢰할 수 있는 분류 결과를 생성하기 위해 두 도메인 영상의 분포가 일치하는 정도를 고려해서 최상의 비지도 도메인 적응 모델을 선택해야 한다.

• 한국국방경영분석학회지
• /
• 제16권2호
• /
• pp.29-42
• /
• 1990

This paper describes a Battle Group Combat Simulation Model (called 'BAGSIM'). BAGSIM is developed to be used as an experimental tool for studies about combat modelling at battle group level. Thus it takes many of the parameters and situations into consideration at this level, and it is designed to be easily adapted to represent equivalent situations to the other more aggregated models. Further the main processes occurring in its simulation procedure such as target detection process, target selection process, firing and killing processes are verified by comparison with the existing stochastic duel models.

• 음성과학
• /
• 제10권4호
• /
• pp.7-21
• /
• 2003

Sums-of-Products models were built for segment duration prediction of spoken Korean. An experiment for the modelling was carried out to apply the results to Korean text-to-speech synthesis systems. 670 read sentences were analyzed. trained and tested for the construction of the duration models. Traditional sequential rule systems were extended to simple additive, multiplicative and additive-multiplicative models based on Sums-of-Products modelling. The parameters used in the modelling include the properties of the target segment and its neighbors and the target segment's position in the prosodic structure. Two optimisation strategies were used: the downhill simplex method and the simulated annealing method. The performance of the models was measured by the correlation coefficient and the root mean squared prediction error (RMSE) between actual and predicted duration in the test data. The best performance was obtained when the data was trained and tested by ' additive-multiplicative models. ' The correlation for the vowel duration prediction was 0.69 and the RMSE. 31.80 ms. while the correlation for the consonant duration prediction was 0.54 and the RMSE. 29.02 ms. The results were not good enough to be applied to the real-time text-to-speech systems. Further investigation of feature interactions is required for the better performance of the Sums-of-Products models.