• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.8
• /
• pp.633-642
• /
• 2016

It is important for advanced active safety systems and autonomous driving cars to get the accurate estimates of the nearby vehicles in order to increase their safety and performance. This paper proposes a sensor fusion method for radar and vision sensors to accurately estimate the state of the preceding vehicles. In particular, we performed a study on compensating for the lateral state error on automotive radar sensors by using a vision sensor. The proposed method is based on the Interactive Multiple Model(IMM) algorithm, which stochastically integrates the multiple Kalman Filters with the multiple models depending on lateral-compensation mode and radar-single sensor mode. In addition, a Probabilistic Data Association Filter(PDAF) is utilized as a data association method to improve the reliability of the estimates under a cluttered radar environment. A two-step correction method is used in the Kalman filter, which efficiently associates both the radar and vision measurements into single state estimates. Finally, the proposed method is validated through off-line simulations using measurements obtained from a field test in an actual road environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.3
• /
• pp.877-893
• /
• 2022

With the development of deep learning, face inpainting has been significantly enhanced in the past few years. Although image inpainting framework integrated with generative adversarial network or attention mechanism enhanced the semantic understanding among facial components, the issues of reconstruction on corrupted regions are still worthy to explore, such as blurred edge structure, excessive smoothness, unreasonable semantic understanding and visual artifacts, etc. To address these issues, we propose a Learnable Structure Knowledge of Fusion Network (LSK-FNet), which learns a prior knowledge by edge generation network for image inpainting. The architecture involves two steps: Firstly, structure information obtained by edge generation network is used as the prior knowledge for face inpainting network. Secondly, both the generated prior knowledge and the incomplete image are fed into the face inpainting network together to get the fusion information. To improve the accuracy of inpainting, both of gated convolution and region normalization are applied in our proposed model. We evaluate our LSK-FNet qualitatively and quantitatively on the CelebA-HQ dataset. The experimental results demonstrate that the edge structure and details of facial images can be improved by using LSK-FNet. Our model surpasses the compared models on L1, PSNR and SSIM metrics. When the masked region is less than 20%, L1 loss reduce by more than 4.3%.

• Journal of Powder Materials
• /
• v.30 no.3
• /
• pp.210-216
• /
• 2023

In this study, machine learning models are proposed to predict the Vickers hardness of AlSi10Mg alloys fabricated by laser powder bed fusion (LPBF). A total of 113 utilizable datasets were collected from the literature. The hyperparameters of the machine-learning models were adjusted to select an accurate predictive model. The random forest regression (RFR) model showed the best performance compared to support vector regression, artificial neural networks, and k-nearest neighbors. The variable importance and prediction mechanisms of the RFR were discussed by Shapley additive explanation (SHAP). Aging time had the greatest influence on the Vickers hardness, followed by solution time, solution temperature, layer thickness, scan speed, power, aging temperature, average particle size, and hatching distance. Detailed prediction mechanisms for RFR are analyzed using SHAP dependence plots.

• ETRI Journal
• /
• v.46 no.2
• /
• pp.277-289
• /
• 2024

In this paper, we describe a neural network-based application that recommends multiple items using dialog context input and simultaneously outputs a response sentence. Further, we describe a multi-item recommendation by specifying it as a set of clothing recommendations. For this, a multimodal fusion approach that can process both cloth-related text and images is required. We also examine achieving the requirements of downstream models using a pretrained language model. Moreover, we propose a gate-based multimodal fusion and multiprompt learning based on a pretrained language model. Specifically, we propose an automatic evaluation technique to solve the one-to-many mapping problem of multi-item recommendations. A fashion-domain multimodal dataset based on Koreans is constructed and tested. Various experimental environment settings are verified using an automatic evaluation method. The results show that our proposed method can be used to obtain confidence scores for multi-item recommendation results, which is different from traditional accuracy evaluation.

• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.7
• /
• pp.637-648
• /
• 2007

This paper presents a sensor fusion method based on indirect Kalman filter(IKF) for error compensation of low-cost inertial sensors and its application to the determination of attitude and position of small flying robots. First, the analysis of the measurement error characteristics to zero input is performed, focusing on the bias due to the temperature variation, to derive a simple nonlinear bias model of low-cost inertial sensors. Moreover, from the experimental results that the coefficients of this bias model possess non-deterministic (stochastic) uncertainties, the bias of low-cost inertial sensors is characterized as consisting of both deterministic and stochastic bias terms. Then, IKF is derived to improve long term stability dominated by the stochastic bias error, fusing low-cost inertial sensor measurements compensated by the deterministic bias model with non-inertial sensor measurement. In addition, in case of using intermittent non-inertial sensor measurements due to the unreliable data link, the upper and lower bounds of the state estimation error covariance matrix of discrete-time IKF are analyzed by solving stochastic algebraic Riccati equation and it is shown that they are dependant on the throughput of the data link and sampling period. To evaluate the performance of proposed method, experimental results of IKF for the attitude determination of a small flying robot are presented in comparison with that of extended Kaman filter which compensates only deterministic bias error model.

• Journal of Digital Convergence
• /
• v.13 no.7
• /
• pp.169-175
• /
• 2015

An effective way to understand the dynamic and time series that follows the passage of time, as valuation is to establish a model to analyze the phenomena of the system. Model of the decision process is efficient clustering information of the total mass of the time series data of the relevant population been collected in a particular number of sub-groups than to look at all a time to an understand of the overall data through each community-specific model determination. In this study, a sub-grouping of the group and the first of the two process model of each cluster by determining, in the following in sub-population characterized by a fusion with heuristic Bayesian clustering techniques proposed a process which can reduce calculation time and cost was confirmed by experiments using actual effectiveness valuation.

• Transactions of the Korean hydrogen and new energy society
• /
• v.24 no.1
• /
• pp.36-43
• /
• 2013

In this paper, a three-dimensional hydrogen desorption model is applied to a thin double-layered annulus ZrCo hydride bed and validated against the temperature evolution data measured by Kang et al. The present model reasonably captures the bed temperature evolution behavior and the 90% hydrogen discharging time. In addition, the performance of thin double-layered annulus bed is evaluated by comparing with a simple cylindrical bed using hydrogen desorption model. This study provides multi-dimensional contours such as temperature and H/M atomic ratio in the metal hydride region. This numerical study provides fundamental understanding during hydrogen desorption process and indicates that efficient design of the metal hydride bed is critical to achieve rapid hydrogen discharging performance. The present three-dimensional hydrogen desorption model is a useful tool for the optimization of bed design and operating conditions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.10
• /
• pp.2768-2787
• /
• 2023

As an important research direction of the application of computer science in the medical field, the automatic generation technology of radiology report has attracted wide attention in the academic community. Because the proportion of normal regions in radiology images is much larger than that of abnormal regions, words describing diseases are often masked by other words, resulting in significant feature loss during the calculation process, which affects the quality of generated reports. In addition, the huge difference between visual features and semantic features causes traditional multi-modal fusion method to fail to generate long narrative structures consisting of multiple sentences, which are required for medical reports. To address these challenges, we propose a multi-feature optimization Transformer (MFOT) for generating radiology reports. In detail, a multi-dimensional mapping attention (MDMA) module is designed to encode the visual grid features from different dimensions to reduce the loss of primary features in the encoding process; a feature pre-fusion (FP) module is constructed to enhance the interaction ability between multi-modal features, so as to generate a reasonably structured radiology report; a detail enhanced attention (DEA) module is proposed to enhance the extraction and utilization of key features and reduce the loss of key features. In conclusion, we evaluate the performance of our proposed model against prevailing mainstream models by utilizing widely-recognized radiology report datasets, namely IU X-Ray and MIMIC-CXR. The experimental outcomes demonstrate that our model achieves SOTA performance on both datasets, compared with the base model, the average improvement of six key indicators is 19.9% and 18.0% respectively. These findings substantiate the efficacy of our model in the domain of automated radiology report generation.

• BMB Reports
• /
• v.48 no.3
• /
• pp.129-130
• /
• 2015

Autophagy is a tightly regulated lysosome-mediated catabolic process in eukaryotes that maintains cellular homeostasis. A distinguishable feature of autophagy is the formation of double- membrane structures, autophagosome, which envelopes the intracellular cargoes and finally degrades them by fusion with lysosomes. So far, many structures of Atg proteins working on the autophagosome formation have been reported, however those involved in autophagosome maturation, a fusion with lysosome, are relatively unknown. One of the molecules in autophagosome maturation, TECPR1, has been identified and recently, structural studies on both ATG5-TECPR1 and ATG5-ATG16L1 complexes revealed that TECPR1 and ATG16L1 share the same binding site on ATG5. These results, in combination with supporting biochemical and cellular biological data, provide an insight into a model for swapping ATG5 partners for autophagosome maturation.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.902-904
• /
• 2003

The main objective of this paper is to investigate the potential utility of multisensor remotely sensed data for improved coastal wetland mapping. Five data fusion models, three algebraic models (Multiplicative (MT), Brovey (BT) and Wavelet transform (WT)) and two spectral domain models (Principals component transform (PCT) and Intensity-Hue-Saturation (IHS)) were implemented and tested over the multisensor data. The fused images were then compared based on visual and statistical approaches. The results show that the wavelet transform provides greater flexibility for combining optical data sets and has good potential for preserving the spatial and spectral content of the original images . However, this model yields poor information when combining optical and microwave data. Brovey transform is more reliable for fusing optical and microwave image data and yields improved information about different wetland features of the coastal zone.