• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.1188-1203
• /
• 2020

The human expression contains a lot of information that can be used to detect complex conditions such as pain and fatigue. After deep learning became the mainstream method, the traditional feature extraction method no longer has advantages. However, in order to achieve higher accuracy, researchers continue to stack the number of layers of the neural network, which makes the real-time performance of the model weak. Therefore, this paper proposed an expression recognition framework based on densely concatenated convolutional neural networks to balance accuracy and latency and apply it to humanoid robots. The techniques of feature reuse and parameter compression in the framework improved the learning ability of the model and greatly reduced the parameters. Experiments showed that the proposed model can reduce tens of times the parameters at the expense of little accuracy.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 1998.03a
• /
• pp.224-230
• /
• 1998

Non-heat-treated type material has been investigated for formability compared to heat-treated material. It is very important in automation of metal forming, since it has difficulties of controlling heat treating system by the computer and has bottle neck problem related with heat treatment. In this paper, we have concerned about mechanical properties of non-heat-treated material after the forging. To compare the characteristics between heat-treated material and non-heat-treated material, tensile, compression and fatigue test has been performed. Considering results of mechanical properites of non-heat-treated material, it can replace heat-treated material. Therefore non-heat-treated material may be applied to cold forging.

• Elastomers and Composites
• /
• v.23 no.4
• /
• pp.299-307
• /
• 1988

The mono-, di- and polysulfidic concentrations of the carbon black filled NR and SBR vucanizates which have different cure systems were determined. The n-decane was used as a swelling solvent. The propane-2-thiol and hexane thiol were used as specific chemical probes to cleavage the chemical crosslinks such as di- and polysulfide. In order to understand the effect of crosslink structure on the physical properties of vulcanizates, the tensile properties, fatigue failure, heat build up, compression set and viscoelastic properties were measured. From the results, the significant relationships between crosslink sturctures and physical properties, especially a gins resistance were obtained. Therefore, the better performances of the product can be satisfied by using the interpretations and approaches in this study.

• International Journal of Highway Engineering
• /
• v.15 no.5
• /
• pp.57-64
• /
• 2013

PURPOSES : It is to evaluate lightweight soil as a subgrade material based on mechanical tests and calculation of pavement performance. METHODS : In this research, various contents of cement and air foam are used to make lightweight soil using wasted dredged soil. Uniaxial compressive strength test is conducted to evaluate strength of 7 and 28 day cured specimens. Secant modulus was calculated based on the stress and strain relationship of uniaxial compressive strength test. Resilient modulus test was measured using by repeated triaxial compression test. The measured resilient modulus was used in layered elastic program to predict fatigue and rutting life at a given pavement structure. RESULTS : Uniaxial compressive strength increases as cement content increases but decrease as air foam content increases. Resilient modulus also increases as cement content increases and decrease as air foam content decrease. CONCLUSIONS : It is concluded that dredge clay soil can be used as subgrade layer material using by lightweight treated soil method.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.2
• /
• pp.181-188
• /
• 2004

The objective of this paper is to numerically examine the mechanical behavior of the swaged power steering(PS) hose subject to internal pressure. PS hose experiences a large internal pressure change in operating, so it's material part has to resist a cyclic expansion and compression without causing oil leakage. This cyclic pressure is intimately associated with fatigue failure of PS hose. In this study, we compare two types of PS hose. The numerical investigation is composed of three steps; swaging analysis, low and high pressure analyses. The comparative numerical results provide the basic data for the optimal PS design.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.14 no.1 s.33
• /
• pp.44-54
• /
• 1990

The purpose of this study was to investigate the effects of thickness of the sleeping pad on the physiological responses and the microclimate in the heating ondol room during sleeping time. The measuring points were skin temp., weight loss and body movement as physiological responses and the compression ratio, the temp. on/under the sleeping pad, the inside limp. of the sleeping quilts and subjective sence while 7 hours sleeping. Thickness of sleeping pads was 27.2 mm(A), 34.8 mm (B), 47.9 mm(C) used for 90 days and no used pad 60.7 mm(D). The floor surface temp. was keeping $29{\pm}1^{\circ}C$ while the environmental conditions was at $23{\pm}1^{\circ}C$ ($50{\pm}5\%$ R.H.) The results were as followings; 1) The mean compression ratio after using the sleeping pads for 45 days was about $70\%$ , and it wasn't increased any more thereafter 2) After 2 hours sleeping the temp. under the sleeping pads was shown that C and D were higher than A and B, and the temp. on the sleeping pads was shown that A and C pads were higher than B and D. But after 3-4 hours sleeping the skin temp. of legs was shown that B and D pads were higher than A and C. 3) Weight loss for C pad was significantly higher than others. Body movement and side position in tying was higher in A and C pads than B and D. 4) The inside temp. of quilt for A and C pads were higher in the early part of sleeping period than those of B and C pads and relative humidity was lower in C than others. 5) D pad was softer than others and fatigue degree was turned out to be lowest in D, the highest in A. From this point of view, we concluded that thickness of the sleeping Pad in the heating ondol room had different characteristic from that in no heating ondol room. And 35 mm used pad was turned out to be proper.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.9 no.3
• /
• pp.125-131
• /
• 1989

Since the paved road suffers from various types of repeated loads for the duration of it's life, it is likely to cause permanent deformation and fatigue finaly destroying the pavement performance. Accordingly, if we are to keep the pavement performance in good condition, it is required to take staps to prevent such troubles from happening in each stage of pavement, and thus to improve the stability of pavement. We find it is quite important to settle the problems such as permanent deformation and fatigue rupture by repeated loads both on subbase course and on subgrade. In this regard, we examined the deformation characteristics of soil cements, on which repeated loads are applied. For the effective examination, we chose to use soil-cements made of cohesive soil and sandy soil respectively, which had $20kg/cm^2$ of unconfined compression strength, at the age of 7 days. The experimental results are: 1. The elastic modulus of soil cement from sandy soil is higher than that of soil cement from cohesive soil. 2. The elastic modulus thends to decrease as the repeated loads rund up to 1,000 times, while increasing between 1,000 times and $1{\times}10^5$ times. 3. Unconfined compression strength is seen to increase about 30%.

• Journal of Chest Surgery
• /
• v.49 no.6
• /
• pp.456-460
• /
• 2016

Background: Surgical correction needs to be considered when diaphragm eventration leads to impaired ventilation and respiratory muscle fatigue. Plication to sufficiently tense the diaphragm by VATS is not as easy to achieve as plication by open surgery. We used pneumatic compression with carbon dioxide ($CO_2$) gas in thoracoscopic diaphragmatic plication and evaluated feasibility and efficacy. Methods: Eleven patients underwent thoracoscopic diaphragmatic plication between January 2008 and December 2013 in Pusan National University Hospital. Medical records were retrospectively reviewed, and compared between the group using $CO_2$ gas and group without using $CO_2$ gas, for operative time, plication technique, duration of hospital stay, postoperative chest tube drainage, pulmonary spirometry, dyspnea score pre- and postoperation, and postoperative recurrence. Results: The improvement of forced expiratory volume at 1 second in the group using $CO_2$ gas and the group not using $CO_2$ gas was $22.46{\pm}11.27$ and $21.08{\pm}5.39$ (p=0.84). The improvement of forced vital capacity 3 months after surgery was $16.74{\pm}10.18$ (with $CO_2$) and $15.6{\pm}0.89$ (without $CO_2$) (p=0.03). During follow-up ($17{\pm}17$ months), there was no dehiscence in plication site and relapse. No complications or hospital mortalities occurred. Conclusion: Thoracoscopic plication under single lung ventilation using $CO_2$ insufflation could be an effective, safe option to flatten the diaphragm.

• Journal of the Korean Society for Railway
• /
• v.16 no.6
• /
• pp.454-459
• /
• 2013

Pressure waves are generated and propagate in a tunnel when train enters tunnel high speed. A compression wave due to the entry of train head propagates along the tunnel and is reflected at tunnel exit as an expansion wave. An expansion wave due to the entry of the train tail propagates along the tunnel and is reflected at tunnel exit as a compression wave. These pressure waves are repeatedly propagated and reflected at the tunnel entrance and exit. Severe pressure changes causes ear-discomfort for passengers in the cabin and micro pressure waves around the tunnel exit. It is necessary to analyze the transient pressure phenomena in tunnels qualitatively and quantitatively, because pressure change rate is considered as one of the major design parameters for optimal tunnel cross sectional area and repeated fatigue force on car body. In this study, we developed a characteristics method based on a fixed mesh system and boundary conditions for crossing trains and analyzed this system using an X-t diagram. The results of the simulation show that offsetting of pressure waves occurs for special entry conditions of a crossing train.

• Journal of the Korean Society for Railway
• /
• v.15 no.5
• /
• pp.436-441
• /
• 2012

Pressure waves are generated and propagate in tunnel when train enters a tunnel with high speed. Compression wave due to the entry of train head propagates along the tunnel and is reflected at tunnel exit as expansion wave. While expansion wave due to the entry of train tail propagates along the tunnel and is reflected at tunnel exit as compression wave. These pressure waves are repeatedly propagated and reflected at tunnel entrance and exit. Severe pressure change per second causes ear-discomfort for passengers in cabin and micro pressure wave around tunnel exit. It is necessary to analyze the transient pressure phenomena in tunnel qualitatively and quantitatively, because pressure change rate is considered as one of major design parameters for an optimal tunnel cross sectional area and the repeated fatigue force on car body. In this study, we developed the characteristics method analysis based on fixed mesh system and compared with the results of real train test. The results of simulation agreed with that of experiment.