Puzzlement Detection from Facial Expression Using Active Appearance Models and Support Vector Machines

Jinwei Wang,Xirong Ma,Jizhou Sun,Ziping Zhao,Yuanping Zhu 보안공학연구지원센터 2014 International Journal of Signal Processing, Image Vol.7 No.5

Affective state detection, as an emerging field of artificial intelligence, is the key to designing effective natural human-computer interaction, especially for e-learning. It will be helpful to make the computer understand learners’ perceptions and provide appropriate guidance, just like teachers in traditional face-to-face classroom learning. Puzzlement is the most frequent non-neutral affective state in learning, and it is usually a sign that learners need more information and guidance. In this paper, we explore a machine learning approach for puzzlement detection from natural facial expression. We use active appearance models (AAMs) to decouple shape and appearance parameters from the face video sequences. Support vector machines (SVMs) are utilized to classify puzzlement and non-puzzlement with several features derived from AAMs. Using a 10-fold cross validation, we achieve the highest recognition rate of 98.9%. Experimental results indicate the feasibility of automatic frame-level puzzlement detection.

Dispersion behavior and thermal conductivity characteristics of Al2O3–H2O nanofluids

Dongsheng Zhu,Xinfang Li,Nan Wang,Xianju Wang,Jinwei Gao,Hua Li 한국물리학회 2009 Current Applied Physics Vol.9 No.1

Nanofluid is a kind of new engineering material consisting of solid nanoparticles with sizes typically of 1–100 nm suspended in base fluids. In this study, Al2O3–H2O nanofluids were synthesized, their dispersion behaviors and thermal conductivity in water were investigated under different pH values and different sodium dodecylbenzenesulfonate (SDBS) concentration. The sedimentation kinetics was determined by examining the absorbency of particle in solution. The zeta potential and particle size of the particles were measured and the Derjaguin–Landau–Verwey–Overbeek (DLVO) theory was used to calculate attractive and repulsive potentials. The thermal conductivity was measured by a hot disk thermal constants analyser. The results showed that the stability and thermal conductivity enhancements of Al2O3–H2O nanofluids are highly dependent on pH values and different SDBS dispersant concentration of nano-suspensions, with an optimal pH value and SDBS concentration for the best dispersion behavior and the highest thermal conductivity. The absolute value of zeta potential and the absorbency of nano-Al2O3 suspensions with SDBS dispersant are higher at pH 8.0. The calculated DLVO interparticle interaction potentials verified the experimental results of the pH effect on the stability behavior. The Al2O3–H2O nanofluids with an ounce of Al2O3 have noticeably higher thermal conductivity than the base fluid without nanoparticles, for Al2O3 nanoparticles at a weight fraction of 0.0015 (0.15 wt%), thermal conductivity was enhanced by up to 10.1%. Nanofluid is a kind of new engineering material consisting of solid nanoparticles with sizes typically of 1–100 nm suspended in base fluids. In this study, Al2O3–H2O nanofluids were synthesized, their dispersion behaviors and thermal conductivity in water were investigated under different pH values and different sodium dodecylbenzenesulfonate (SDBS) concentration. The sedimentation kinetics was determined by examining the absorbency of particle in solution. The zeta potential and particle size of the particles were measured and the Derjaguin–Landau–Verwey–Overbeek (DLVO) theory was used to calculate attractive and repulsive potentials. The thermal conductivity was measured by a hot disk thermal constants analyser. The results showed that the stability and thermal conductivity enhancements of Al2O3–H2O nanofluids are highly dependent on pH values and different SDBS dispersant concentration of nano-suspensions, with an optimal pH value and SDBS concentration for the best dispersion behavior and the highest thermal conductivity. The absolute value of zeta potential and the absorbency of nano-Al2O3 suspensions with SDBS dispersant are higher at pH 8.0. The calculated DLVO interparticle interaction potentials verified the experimental results of the pH effect on the stability behavior. The Al2O3–H2O nanofluids with an ounce of Al2O3 have noticeably higher thermal conductivity than the base fluid without nanoparticles, for Al2O3 nanoparticles at a weight fraction of 0.0015 (0.15 wt%), thermal conductivity was enhanced by up to 10.1%.

Modification of tool influence function for bonnet polishing tool based on analysis of interfacial contact state

Ri Pan,Xiangxiang Zhu,Zhenzhong Wang,Dongju Chen,Shuting Ji,Jinwei Fan,Rui Wang 대한기계학회 2022 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.36 No.6

The polishing mechanism of bonnet polishing (BP) and the tool influence function (TIF) of bonnet tool have been widely studied. However, most of current studies pay little attention to the influence of polishing slurry on the BP mechanism as well as TIF. This study proposes that the interfacial contact state between the polishing tool and the workpiece is in a mixed lubrication state, subsequently the BP mechanism is further explored. It is found that the workpiece material is removed by polishing pad and abrasives. The polishing slurry is not directly involved in workpiece removal, but shares the normal pressure of the polishing tool on workpiece, which affects material removal. Based on the above mechanism, the TIF removal prediction model is established and verified by experiments. The results show that the maximum error of the model prediction value is quiet small, which verifies the model. Moreover, compared with Preston model and the previous model, which ignored the influence of the fluid, the average prediction error of the model in this paper when D 0 = 20 mm is 6.38 %, while the previous model and Preston model are 11.21 % and 49.10 %, respectively. Which illustrates the model in this paper has higher accuracy.

Numerical investigation of effects of rotating downdraft on tornado-like-vortex characteristics

Shuyang Cao,Mengen Wang,Jinwei Zhu,Jinxin Cao,Tetsuro Tamura,Qingshan Yang 한국풍공학회 2018 Wind and Structures, An International Journal (WAS Vol.26 No.3

Appropriate modeling of a tornado-like vortex is a prerequisite when studying the near-ground wind characteristics of a tornado and tornado-induced wind loads on structures. Both Ward- and ISU-type tornado simulators employ guide vanes to induce angular momentum to converge flow in order to generate tornado-like vortices. But in the Ward-type simulator, the guide vanes are mounted near the ground while in the ISU-type they are located at a high position to allow vertical circulation of flow that creates a rotating downdraft to generate a tornado-like vortex. In this study, numerical simulations were performed to reproduce tornado-like vortices using both Ward-type and ISU-type tornado simulators, from which the effects of rotating downdraft on the vortex characteristics were clarified. Particular attention was devoted to the wander of tornado-like vortices, and their dependences on swirl ratio and fetch length were investigated. The present study showed that the dynamic vortex structure depends significantly on the vortex-generating mechanism, although the time- averaged structure remains similar. This feature should be taken into consideration when tornado-like-vortex simulators are utilized to investigate tornado-induced wind forces on structures.

Time Series Data Cleaning Method Based on Optimized ELM Prediction Constraints

Guohui Ding,Yueyi Zhu,Chenyang Li,Jinwei Wang,Ru Wei,Zhaoyu Liu 한국정보처리학회 2023 Journal of information processing systems Vol.19 No.2

Affected by external factors, errors in time series data collected by sensors are common. Using the traditionalmethod of constraining the speed change rate to clean the errors can get good performance. However, they areonly limited to the data of stable changing speed because of fixed constraint rules. Actually, data with unevenchanging speed is common in practice. To solve this problem, an online cleaning algorithm for time series databased on dynamic speed change rate constraints is proposed in this paper. Since time series data usually changesperiodically, we use the extreme learning machine to learn the law of speed changes from past data and predictthe speed ranges that change over time to detect the data. In order to realize online data repair, a dual-windowmechanism is proposed to transform the global optimal into the local optimal, and the traditional minimumchange principle and median theorem are applied in the selection of the repair strategy. Aiming at the problemthat the repair method based on the minimum change principle cannot correct consecutive abnormal points,through quantitative analysis, it is believed that the repair strategy should be the boundary of the repaircandidate set. The experimental results obtained on the dataset show that the method proposed in this paper canget a better repair effect.

Numerical investigation of effects of rotating downdraft on tornado-like-vortex characteristics

Cao, Shuyang,Wang, Mengen,Zhu, Jinwei,Cao, Jinxin,Tamura, Tetsuro,Yang, Qingshan Techno-Press 2018 Wind and Structures, An International Journal (WAS Vol.26 No.3

Appropriate modeling of a tornado-like vortex is a prerequisite when studying the near-ground wind characteristics of a tornado and tornado-induced wind loads on structures. Both Ward- and ISU-type tornado simulators employ guide vanes to induce angular momentum to converge flow in order to generate tornado-like vortices. But in the Ward-type simulator, the guide vanes are mounted near the ground while in the ISU-type they are located at a high position to allow vertical circulation of flow that creates a rotating downdraft to generate a tornado-like vortex. In this study, numerical simulations were performed to reproduce tornado-like vortices using both Ward-type and ISU-type tornado simulators, from which the effects of rotating downdraft on the vortex characteristics were clarified. Particular attention was devoted to the wander of tornado-like vortices, and their dependences on swirl ratio and fetch length were investigated. The present study showed that the dynamic vortex structure depends significantly on the vortex-generating mechanism, although the time-averaged structure remains similar. This feature should be taken into consideration when tornado-like-vortex simulators are utilized to investigate tornado-induced wind forces on structures.

Tuning of BixOyCl formation with sonication time during ultrasound-hydrothermal preparation

Zeqing Long,Guangming Zhang,Ting Wei,Lijun Niu,Jia Zhu,Jinwei Li 한국공업화학회 2020 Journal of Industrial and Engineering Chemistry Vol.84 No.-

Developing novel and visible-light drive photocatalysts is a hot topic, and bismuth oxychloride is one ofthe popular photocatalysts. In this paper, a series of bismuth oxychloride were synthesized by ultrasonic-hydrothermal method with different ultrasonic time. By simply changing the ultrasonic time, we cancontrol the morphology, structure, stoichiometry and photoelectric performance of bismuth oxychloride. The possible mechanism for the formation of BixOyCl photocatalysts involved the ultrasonic destructionof chemical bonds and increase of oxygen vacancy concentration, and the influence of ultrasonic on thepH during preparation. The sample of ultrasonic 30 min (U-30) showed the best photocatalytic activitydue to the suitable morphology, structure, photoelectric performance, the formation of theheterostructure, and the presence of oxygen vacancies. For the removal of Rhodamine B (/Ciprofloxacin),the reaction rate constant for U-30 was 10.4 (/3.6), 4.9 (/1.6), 6.9 (/4.8), and 9.3 (/2.3) times that of the U-0,U-10, U-20, and U-40, respectively. Photoluminescence spectra showed that the photogeneratedelectron-hole recombination rate decreased gradually with the ultrasonic time. Further, the possiblemechanism of photocatalytic process was investigated. This research provides a green, economical andsimple method to modify bismuth oxychloride.

Kaempferol attenuates spinal cord injury by interfering inflammatory and oxidative stress by targeting the p53 protein: a molecular docking analysis

Chen Lu,Cao Kai,Gu Yurong,Luo Chao,Mao Wei,Zhou Weijun,Zhu Jinwei,Zhang Huying 대한독성 유전단백체 학회 2021 Molecular & cellular toxicology Vol.17 No.3

Background Kaempferol (KMF) is a fl avone reported to have anti-oxidant and anti-infl ammatory activity. Objective The present study screened the eff ect of KMF in the animal model of spinal cord injury (SCI). Results KMF caused a signifi cant inhibition of spinal cord injury mediated oxidative stress and also suppressed the infl ammatory reactions. The treatment of KMF also inhibited the levels of p53, TGF-β1 and COX-2 whereas a signifi cant elevation in Bcl-2/Bax ratio was observed after the rats were treated with KMF. The in silico docking analysis suggested potential binding of KMF having lower energy with p53 confi rming the potential target of KMF. Conclusion The treatment of KMF exerted neuroprotective eff ect by improving the anti-oxidant status and infl ammatory response. The signifi cant spinal cord injury protective eff ect of KMF in rats was attributed by targeting p53 and Bcl-2/Bax ratio.