• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.1
• /
• pp.35-45
• /
• 2007

The log data collected from mobile devices contains diverse meaningful and practical personal information. However, this information is usually ignored because of its limitation of memory capacity, computation power and analysis. We propose a novel method that detects landmarks of meaningful information for users by analyzing the log data in distributed modules to overcome the problems of mobile environment. The proposed method adopts Bayesian probabilistic approach to enhance the inference accuracy under the uncertain environments. The new cooperative modularization technique divides Bayesian network into modules to compute efficiently with limited resources. Experiments with artificial data and real data indicate that the result with artificial data is amount to about 84% precision rate and about 76% recall rate, and that including partial matching with real data is about 89% hitting rate.

• Journal of Korea Society of Industrial Information Systems
• /
• v.25 no.4
• /
• pp.9-16
• /
• 2020

Pronunciation recognition systems that use only video information and ignore voice information can be applied to various customized services. In this paper, we develop a system that applies a Bayesian classifier to distinguish Korean vowels via lip shapes in images. We extract feature vectors from the lip shapes of facial images and apply them to the designed machine learning model. Our experiments show that the system's recognition rate is 94% for the pronunciation of 'A', and the system's average recognition rate is approximately 84%, which is higher than that of the CNN tested for comparison. Our results show that our Bayesian classification method with feature values from lip region landmarks is efficient on a small training set. Therefore, it can be used for application development on limited hardware such as mobile devices.

• Journal of Digital Convergence
• /
• v.17 no.5
• /
• pp.347-351
• /
• 2019

The purpose of this study is to suggest the Assessment model for tourist information center in Seoul. As a research method, we analyzed international guideline and interview with tourism experts in order to rate the tourist centers in Seoul. Secondly, we renamed the international rating model to Itaewon information center as a typical landmark in Seoul. The assessment factors for T.I.C is assembled through researching of the centers' status in terms of overall service satisfaction. Via in - depth interview with 9 visitors, as a result, we were able to derive the possibility that new-designed rating model is able to be applied to the Tourist centers in Seoul. It is significant that this study suggests ways to improve domestic tourist center service. It is expected that the follow - up study will help improve the factors to Seoul T.I.C, not only Itaewon, with much more specific rating method.

• 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%.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.4
• /
• pp.187-194
• /
• 2016

When performing experiments in indoor and outdoor environment, we need a system that monitors a volunteer to prevent dangerous situations and efficiently manages the data in real time. We developed a guidance device for visually impaired person that guides the user to walk safely to the destination in the previous study. We set a POI (Point of Interest) of a specific location indoors and outdoors and tracks the user's position and navigate the walking path using artificial markers and ZigBee modules as landmark. In addition, we develop path finding algorithm to be used for navigation in the guidance device. In the test bed, the volunteers are exposed to dangerous situations and can be an accident due to malfunction of the device since they are visually impaired person or normal person wearing a eye patch. Therefore the device requires a system that remotely monitors the volunteer wearing guidance device and manages indoor or outdoor a lot of map data. In this paper, we introduce a managing system that monitors the volunteers remotely and handles map data efficiently. We implement a management system which can monitor the volunteer in order to prevent a hazardous situation and effectively manage large amounts of data. In addition, we verified the effectiveness of the proposed system through various experiments.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.9
• /
• pp.25-31
• /
• 2020

In this study, for the safe use of a portable ultrasonic skin-beauty device, an android app was developed to show the irradiation locations of focused ultrasound to a user through augmented reality (AR) and enable stable self-surgery. The utility of the app was assessed through testing. While the user is making a facial treatment with the beauty device, the user's face and the ultrasonic irradiation location on the face are detected in real-time with a smart-phone camera. The irradiation location is then indicated on the face image and shown to the user so that excessive ultrasound is not irradiated to the same area during treatment. To this end, ML-Kit is used to detect the user's face landmarks in real-time, and they are compared with a reference face model to estimate the pose of the face, such as rotation and movement. After mounting a LED on the ultrasonic irradiation part of the device and operating the LED during irradiation, the LED light was searched to find the position of the ultrasonic irradiation on the smart-phone screen, and the irradiation position was registered and displayed on the face image based on the estimated face pose. Each task performed in the app was implemented through the thread and the timer, and all tasks were executed within 75 ms. The test results showed that the time taken to register and display 120 ultrasound irradiation positions was less than 25ms, and the display accuracy was within 20mm when the face did not rotate significantly.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.3
• /
• pp.99-108
• /
• 2021

There are unrestricted conditions on the driver's face inside the vehicle, such as changes in lighting, partial occlusion and various changes in the driver's condition. In this paper, we propose a face identification system in an unrestricted vehicle environment. The proposed method uses a near-infrared (NIR) camera to minimize the changes in facial images that occur according to the illumination changes inside and outside the vehicle. In order to process a face exposed to extreme light, the normal face image is changed to a simulated overexposed image using mean and variance for training. Thus, facial classifiers are simultaneously generated under both normal and extreme illumination conditions. Our method identifies a face by detecting facial landmarks and aggregating the confidence score of each landmark for the final decision. In particular, the performance improvement is the highest in the class where the driver wears glasses or sunglasses, owing to the robustness to partial occlusions by recognizing each landmark. We can recognize the driver by using the scores of remaining visible landmarks. We also propose a novel robust rejection and a new evaluation method, which considers the relations between registered and unregistered drivers. The experimental results on our dataset, PolyU and ORL datasets demonstrate the effectiveness of the proposed method.

• Fashion & Textile Research Journal
• /
• v.18 no.4
• /
• pp.468-476
• /
• 2016

Recent advancements in optics technology enable us to realize fast scans of hands using two-dimensional (2D) image scanners. In this paper, we propose an automatic hand measurement system using 2D image scanners for customized glove production. To develop the automatic hand measurement system, firstly hand scanning devices has been constructed. The devices are designed to block external lights and have user interface to guide hand posture during scanning. After hands are scanned, hand contour is extracted using binary image processing, noise elimination and outline tracing. And then, 19 hand landmarks are automatically detected using an automatic hand landmark detection algorithm based on geometric feature analysis. Then, automatic hand measurement program is executed based on the automatically extracted landmarks and measurement algorithms. The automatic hand measurement algorithms have been developed for 18 hand measurements required for custom-made glove pattern making. The program has been coded using the C++ programming language. We have implemented experiments to demonstrate the validity of the system using 11 subjects (8 males, 3 females) by comparing automatic 2D scan measurements with manual measurements. The result shows that the automatic 2D scan measurements are acceptable in the customized glove making industry. Our evaluation results confirm its effectiveness and robustness.

• Journal of Communications and Networks
• /
• v.14 no.5
• /
• pp.546-553
• /
• 2012

Compact routing intends to achieve good tradeoff between the routing path length and the memory overhead, and is recently considered as a main alternative to overcome the fundamental scaling problems of the Internet routing system. Plenty of studies have been conducted on compact routing, and quite a few universal compact routing schemes have been designed for arbitrary network topologies. However, it is generally believed that specialized compact routing schemes for peculiar network topologies can have better performance than universal ones. Studies on complex networks have uncovered that most real-world networks exhibit power-law degree distributions, i.e., a few nodes have very high degrees while many other nodes have low degrees. High-degree nodes play the crucial role of hubs in communication and inter-networking. Based on this fact, we propose two highest-degree landmark based compact routing schemes, namely HDLR and $HDLR^+$. Theoretical analysis on random power-law graphs shows that the two schemes can achieve better space-stretch trade-offs than previous compact routing schemes. Simulations conducted on random power-law graphs and real-world AS-level Internet graph validate the effectiveness of our schemes.

• Communications for Statistical Applications and Methods
• /
• v.30 no.1
• /
• pp.37-48
• /
• 2023

Shape analysis is a method for measuring, describing and comparing the shape of objects in geometric space. An important aspect is to obtain Procrustes distance based on least square method. We note that the shape is all the geometrical information that remains when location, scale and rotational effects are filtered out from an object. However, and unfortunately, when we cannot measure some landmarks which are some biologically or geometrically meaningful points of any object, it is not possible to measure the variation of all shapes of an object, including that of the incomplete object. Hence, we need to replace the missing landmarks. In particular, Albers and Gower (2010) studied the missing rows of configurations in Procrustes analysis. They noted that the convergence of their approach can be quite slow. In this study, alternatively, we derive an algorithm for estimating the missing landmarks based on the pre-shapes. The pre-shape is invariant under the location and scaling of the original configuration with the centroid size of the pre-shape being one. Therefore we expect that we can reduce the amount of total computing time for obtaining the estimate of the missing landmarks.