• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.313-321
• /
• 2017

One of the most frequently performed tasks in human-robot interaction (HRI), intelligent vehicles, and security systems is face related applications such as face recognition, facial expression recognition, driver state monitoring, and gaze estimation. In these applications, accurate head pose estimation is an important issue. However, conventional methods have been lacking in accuracy, robustness or processing speed in practical use. In this paper, we propose a novel method for estimating head pose with a monocular camera. The proposed algorithm is based on a deep neural network for multi-task learning using a small grayscale image. This network jointly detects multi-view faces and estimates head pose in hard environmental conditions such as illumination change and large pose change. The proposed framework quantitatively and qualitatively outperforms the state-of-the-art method with an average head pose mean error of less than $4.5^{\circ}$ in real-time.

• Journal of Korea Multimedia Society
• /
• v.15 no.4
• /
• pp.449-471
• /
• 2012

In intelligent surveillance systems, it is required to robustly track multiple people. Most of the previous studies adopted a Gaussian mixture model (GMM) for discriminating the object from the background. However, it has a weakness that its performance is affected by illumination variations and shadow regions can be merged with the object. And when two foreground objects overlap, the GMM method cannot correctly discriminate the occluded regions. To overcome these problems, we propose a new method of tracking and identifying multiple people. The proposed research is novel in the following three ways compared to previous research: First, the illuminative variations and shadow regions are reduced by an illumination normalization based on the median and inverse filtering of the L*a*b* image. Second, the multiple occluded and overlapped people are tracked by combining the GMM in the still image and the Lucas-Kanade-Tomasi (LKT) method in successive images. Third, with the proposed human tracking and the existing face detection & recognition methods, the tracked multiple people are successfully identified. The experimental results show that the proposed method could track and recognize multiple people with accuracy.

• The Journal of Korea Robotics Society
• /
• v.8 no.4
• /
• pp.266-272
• /
• 2013

Facial feature extraction and tracking are essential steps in human-robot-interaction (HRI) field such as face recognition, gaze estimation, and emotion recognition. Active shape model (ASM) is one of the successful generative models that extract the facial features. However, applying only ASM is not adequate for modeling a face in actual applications, because positions of facial features are unstably extracted due to limitation of the number of iterations in the ASM fitting algorithm. The unaccurate positions of facial features decrease the performance of the emotion recognition. In this paper, we propose real-time facial feature extraction and tracking framework using ASM and LK optical flow for emotion recognition. LK optical flow is desirable to estimate time-varying geometric parameters in sequential face images. In addition, we introduce a straightforward method to avoid tracking failure caused by partial occlusions that can be a serious problem for tracking based algorithm. Emotion recognition experiments with k-NN and SVM classifier shows over 95% classification accuracy for three emotions: "joy", "anger", and "disgust".

• Journal of KIISE:Software and Applications
• /
• v.30 no.3_4
• /
• pp.372-378
• /
• 2003

Limitations on the linear discriminant analysis (LDA) for face rerognition, such as the loss of generalization and the computational infeasibility, are addressed and illustrated for a small number of samples. The principal component analysis (PCA) followed by the LDA mapping may be an alternative that ran overcome these limitations. We also show that any schemes based on either mappings or template matching are vulnerable to image variations due to rotation, translation, facial expressions, or local illumination conditions. This entails the importance of a proper preprocessing that can compensate for such variations. A simple template matching, when combined with the geometrically correlated feature-based detection as a preprocessing, is shown to outperform mapping techniques in terms of both the accuracy and the robustness to image variations.

• Structural Engineering and Mechanics
• /
• v.57 no.3
• /
• pp.473-506
• /
• 2016

The aim of this work is the development of a 2D quadrilateral isoparametric finite element model, based on a layerwise approach, for the bending analysis of sandwich plates. The face sheets and the core are modeled individually using, respectively, the first order shear deformation theory and the third-order plate theory. The displacement continuity condition at the interfaces 'face sheets-core' is satisfied. The assumed natural strains method is introduced to avoid an eventual shear locking phenomenon. The developed element is a four-nodded isoparametric element with fifty two degrees-of-freedom (52 DOF). Each face sheet has only two rotational DOF per node and the core has nine DOF per node: six rotational degrees and three translation components which are common for the all sandwich layers. The performance of the proposed element model is assessed by six examples, considering symmetric/unsymmetric composite sandwich plates with different aspect ratios, loadings and boundary conditions. The numerical results obtained are compared with the analytical solutions and the numerical results obtained by other authors. The results indicate that the proposed element model is promising in terms of the accuracy and the convergence speed for both thin and thick plates.

• The Journal of the Korea Contents Association
• /
• v.6 no.3
• /
• pp.125-134
• /
• 2006

The purpose of this paper is constructed a HEEAAS(Human Emotional Expression Animaion System), which is an animation system to show both the face and the body motion from the inputted voice about just 4 types of emotions such as fear, dislike, surprise and normal. To implement our paper, we chose the korean young man in his twenties who was to show appropriate emotions the most correctly. Also, we have focused on reducing the processing time about making the real animation in making both face and body codes of emotions from the inputted voice signal. That is, we can reduce the search time to use the binary search technique from the face and body motion databases, Throughout the experiment, we have a 99.9% accuracy of the real emotional expression in the cartoon animation.

• Korean Journal of Optics and Photonics
• /
• v.12 no.4
• /
• pp.334-338
• /
• 2001

A Fizeau interferometer without beam splitter was constructed. Single-mode optical fiber was used as a spherical wave source and the face of fiber end was polished and coated to be a reflecting surface. The reflecting surface was angled so that interference fringe could be detected by CCD camera. Beam splitter in front of a spherical wave source could distort the wave front and that was one of the component error sources. With the proposed configuration there was no need to place beam splitter in the system. Improvement of phase measuring accuracy was evaluated quantitatively by comparing the result of this setup with that of a conventional Fizeau interferometer. Wave front of the angled end-face optical fiber source was also measured to verify its sphericity by PS/PDI (Phase Shifting/Point Diffraction Interferometer). The principle of this technique was presented and the experimental results and its applications were discussed. ussed.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.7
• /
• pp.117-125
• /
• 2011

In this paper, it was segmented the face color distribution using YCbCr color model, which is one of the feature-based methods, and preprocessing stage was to be insensitive to the sensitivity for light which is one of the disadvantages for the feature-based methods by the quantization. In addition, it has raised the accuracy of image synthesis with characteristics which is selected the object of the most same image as the shape of pattern using rough set. In this paper, the detection rates of the proposed face detection algorithm was confirmed to be better about 2~3% than the conventional algorithms regardless of the size and direction on the various faces by simulation.

• Advances in nano research
• /
• v.14 no.3
• /
• pp.261-271
• /
• 2023

The present study follows three main goals. First, an analytical solution with high accuracy is developed to assess the effects of embedding pre-strained shape memory alloy (SMA) wires on the critical buckling temperatures of rectangular sandwich plates made of soft core and graphite fiber/epoxy (GF/EP) face sheets based on piecewise low-order shear deformation theory (PLSDT) using Brinson's model. As the second goal, this study compares the effects of SMAs on the thermal buckling of sandwich plates with those of carbon nanotubes (CNTs). The glass transition temperature is considered as a limiting factor. For each material, the effective ranges of operating temperature and thickness ratio are determined for real situations. The results indicate that depending on the geometric parameters and thermal conditions, one of the SMAs and CNTs may outperform the other. The third purpose is to study the thermal buckling of sandwich plates with advanced hybrid SMA/CNT/GF/EP composite face sheets. It is shown that in some circumstances, the co-incorporation of SMAs and CNTs leads to an astonishing enhancement in the critical buckling temperatures of sandwich plates.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.4 no.2
• /
• pp.208-223
• /
• 1994

The purpose of this study was to evaluate the efficiency of diffusive(or passive) sampler in measuring airbone organic solvents. Diffusive samplers are generally simple in construction and do not require power for operation. The efficiency of the diffusive samplers has not sufficiently been investigated in Korea. Three types of samplers were studied in this study. The sampling and analytical results by passive samplers were compared with results by charcoal tube method recommended by NIOSH(National Institute for Occupational Safty and Health). The following characteristics are identified and studied as critical to the performance passive monitors; recovery, reverse diffusion, storage stability, accuracy and precision, face velocity and humidity, n-Hexane, TCE(trichloroethylene) and toluene were used as test vapors. A dynamic vapor exposure system consisting of organic vapor generator and sampling chamber for evaluating diffusive samplers are made. The results of the study are summarized as follows. 1. NIOSH recommands that the overall accuracy of a sampling method in the range of 0.5 to 2.0 times the occupational health standard should be ${\pm}25$ percent for 95 percent confidence level. Among three types of diffusive samplers, sampler A has permeation membrane and samplers Band C have diffusive areas, samplers A and B met the criterion that overall accuracy for 95% confidence level of the samplers were within ${\pm}25$ percent of the reference value. Sampler C had overall accuracy ${\pm}9.6%$ and ${\pm}11.8%$ in hexane and TCE, respectively. The concentration of toluene was overestimated in sampler C with overall accuracy of ${\pm}43.9%$. 2. The desorption efficiencies of diffusive samplers were 96-107%. 3. There was no significant sampe loss during four weeks of storage both with and without refrigeration. 4. There was no significant reverse diffusion, when the samplers were exposure to clean air for 2 hours after sampling for 2 hours at the level of 2 TLY. 5. In case of 8 hours sampling, relative differences(RD) of concentrations between charcoal tube method and diffusive method were 15-39%, 13-46%, and 4-35% for sampler A, B and C, respectively. The performance was poor in 8 hours sampling for multiple substance monitors. 6. At high velocity(100 cm/sec), samplers B and C overestimated the concentrations of organic vapors, and sampler A with permeation membrance gave better results. 7. At 80% relative humidity, samplers showed no siginificant effect. Low humidity also did not affect the diffusive samplers.