• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.1289-1292
• /
• 1998

HCI(human computer interface) technologies have been often implemented using mouse, keyboard and joystick. Because mouse and keyboard are used only in limited situation, More natural HCI methods such as speech based method and gesture based method recently attract wide attention. In this paper, we present multi-modal input system to control Windows system for practical use of multi-media computer. Our multi-modal input system consists of three parts. First one is virtual-hand mouse part. This part is to replace mouse control with a set of gestures. Second one is Windows control system using speech recognition. Third one is Windows control system using gesture recognition. We introduce neural network and HMM methods to recognize speeches and gestures. The results of three parts interface directly to CPU and through Windows.

• Proceedings of the KSME Conference
• /
• 2007.05a
• /
• pp.1590-1594
• /
• 2007

This paper describes a case study on dynamic characteristics analysis of a 5-axis multi-tasking machine tool of ram-head typed. Natural frequency and corresponding vibration modes of the machine tool structure were obtained by using both FEM modal analysis and an experimental modal test(impulse hammer test). Both the theoretical and experiment analysis results showed good agreement with each other. Finally, some discussion and review, from the view point of resonance vibration and/or mode coupled chatter, were made based on the analysis results.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.17 no.7 s.124
• /
• pp.579-586
• /
• 2007

Analytical method to analyze the effect of tolerance on the modal characteristic of multi-body systems in dynamic equilibrium position is suggested in this paper. Monte-Carlo method is conventionally employed to perform the tolerance analysis. However, Monte-Carlo method spends too much time for analysis and has a greater or less accuracy depending on sample condition. To resolve these problems, an analytical method is suggested in this paper. Sensitivity equations for damped natural frequencies and the transfer function are derived at the dynamic equilibrium position. By employing the sensitivity information of mass, damping and stiffness matrices, the sensitivities of damped natural frequencies and the transfer function can be calculated.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.4
• /
• pp.1090-1105
• /
• 2012

A multi-modal sensing M2M healthcare monitoring system for the continuous monitoring of patients under their natural physiological states or elderly persons with chronic diseases is summarized. The system is designed for homecare or the monitoring of the elderly who live in country side or small rest home without enough support from caregivers or doctors, instead of patient monitoring in big hospital environment. Further insights into the natural cause and progression of diseases are afforded by context-aware sensing, which includes the use of accelerometers to monitor patient activities, or by location-aware indoor tracking based on ultrasonic and RF sensing. Moreover, indoor location tracking provides information about the location of patients in their physical environment and helps the caregiver in the provision of appropriate support.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.6
• /
• pp.619-624
• /
• 2014

This paper introduces a robot-based learning system which is designed to teach multiplication to children. In addition to a small humanoid and a smart device delivering educational content, we employ a type of mixed-initiative operation which provides enhanced multi-modal cognition to the r-learning system through human intervention. To investigate major factors that influence people's intention to use the r-learning system and to see how the multi-modality affects the connections, we performed a user study based on TAM (Technology Acceptance Model). The results support the fact that the quality of the system and the natural interaction are key factors for the r-learning system to be used, and they also reveal very interesting implications related to the human behaviors.

• Archives of design research
• /
• v.18 no.2 s.60
• /
• pp.273-282
• /
• 2005

In this study, we proposed Context Cube as a conceptual model of user context, and a Multi-modal Interaction Design framework to develop ubiquitous service through understanding user context and analyzing correlation between context awareness and multi-modality, which are to help infer the meaning of context and offer services to meet user needs. And we developed a case study to verify Context Cube's validity and proposed interaction design framework to derive personalized ubiquitous service. We could understand context awareness as information properties which consists of basic activity, location of a user and devices(environment), time, and daily schedule of a user. And it enables us to construct three-dimensional conceptual model, Context Cube. Also, we developed ubiquitous interaction design process which encloses multi-modal interaction design by studying the features of user interaction presented on Context Cube.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.351-360
• /
• 2018

Human emotion recognition is a research topic that is receiving continuous attention in computer vision and artificial intelligence domains. This paper proposes a method for classifying human emotions through multiple neural networks based on multi-modal signals which consist of image, landmark, and audio in a wild environment. The proposed method has the following features. First, the learning performance of the image-based network is greatly improved by employing both multi-task learning and semi-supervised learning using the spatio-temporal characteristic of videos. Second, a model for converting 1-dimensional (1D) landmark information of face into two-dimensional (2D) images, is newly proposed, and a CNN-LSTM network based on the model is proposed for better emotion recognition. Third, based on an observation that audio signals are often very effective for specific emotions, we propose an audio deep learning mechanism robust to the specific emotions. Finally, so-called emotion adaptive fusion is applied to enable synergy of multiple networks. The proposed network improves emotion classification performance by appropriately integrating existing supervised learning and semi-supervised learning networks. In the fifth attempt on the given test set in the EmotiW2017 challenge, the proposed method achieved a classification accuracy of 57.12%.

• Structural Engineering and Mechanics
• /
• v.25 no.1
• /
• pp.53-73
• /
• 2007

In this study, an effective load increment method for multi modal adaptive non-linear static (pushover) analysis (NSA) for building type structures is presented. In the method, lumped plastisicity approach is adopted and geometrical non-linearties (second-order effects) are included. Non-linear yield conditions of column elements and geometrical non-linearity effects between successive plastic sections are linearized. Thus, load increment needed for formation of plastic sections can be determined directly (without applying iteration or step-by-step techniques) by using linearized yield conditions. After formation of each plastic section, the higher mode effects are considered by utilizing the essentials of traditional response spectrum analysis at linearized regions between plastic sections. Changing dynamic properties due to plastification in the system are used on the calculation of modal lateral loads. Thus, the effects of stiffness changes and local mechanism at the system on lateral load distribution are included. By using the proposed method, solution can be obtained effectively for multi-mode whereby the properties change due to plastifications in the system. In the study, a new procedure for determination of modal lateral loads is also proposed. In order to evaluate the proposed method, a 20 story RC frame building is analyzed and compared with Non-linear Dynamic Analysis (NDA) results and FEMA 356 Non-linear Static Analysis (NSA) procedures using fixed loads distributions (first mode, SRSS and uniform distribution) in terms of different parameters. Second-order effects on response quantities and periods are also investigated. When the NDA results are taken as reference, it is seen that proposed method yield generally better results than all FEMA 356 procedures for all investigated response quantities.

• Design & Manufacturing
• /
• v.15 no.1
• /
• pp.66-71
• /
• 2021

In this study, die turning injection(DTI) mold design for manufacturing reservoir fluid tanks used for cooling in-vehicle batteries, inverters, and motors was conducted based on multi-field CAE. Part design, performance evaluation, and mold design of the reservoir fluid tank was performed. The frequency response characteristics through modal and harmonic response analysis to satisfy the automotive performance test items for the designed part were examined. Analysis of re-melting characteristics and structural analysis of the driving part for designing the rotating die of the DTI mold were performed. Part design was possible when the natural frequency performance value of 32Hz or higher was satisfied through finite element analysis, and the temperature distribution and deformation characteristics of the part after injection molding were found through the first injection molding analysis. In addition, it can be seen that the temperature change of the primary part greatly influences the re-melting characteristics during the secondary injection. The minimum force for driving the turning die of the designed mold was calculated through structural analysis. Hydraulic system design was possible. Finally, a precise and efficient DTI mold design for the reservoir fluid tank was possible through presented multi-field CAE process.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.670-684
• /
• 2024

3D Cross-Modal Retrieval (3DCMR) is a task that retrieves 3D objects regardless of modalities, such as images, meshes, and point clouds. One of the most prominent methods used for 3DCMR is the Cross-Modal Center Loss Function (CLF) which applies the conventional center loss strategy for 3D cross-modal search and retrieval. Since CLF is based on center loss, the center features in CLF are also susceptible to subtle changes in hyperparameters and external inferences. For instance, performance degradation is observed when the batch size is too small. Furthermore, the Mean Squared Error (MSE) used in CLF is unable to adapt to changes in batch size and is vulnerable to data variations that occur during actual inference due to the use of simple Euclidean distance between multi-modal features. To address the problems that arise from small batch training, we propose a Noisy Center Loss (NCL) method to estimate the optimal center features. In addition, we apply the simple Siamese representation learning method (SimSiam) during optimal center feature estimation to compare projected features, making the proposed method robust to changes in batch size and variations in data. As a result, the proposed approach demonstrates improved performance in ModelNet40 dataset compared to the conventional methods.