• ETRI Journal
• /
• v.42 no.1
• /
• pp.78-89
• /
• 2020

Human activity recognition (HAR) has become effective as a computer vision tool for video surveillance systems. In this paper, a novel biometric system that can detect human activities in 3D space is proposed. In order to implement HAR, joint angles obtained using an RGB-depth sensor are used as features. Because HAR is operated in the time domain, angle information is stored using the sliding kernel method. Haar-wavelet transform (HWT) is applied to preserve the information of the features before reducing the data dimension. Dimension reduction using an averaging algorithm is also applied to decrease the computational cost, which provides faster performance while maintaining high accuracy. Before the classification, a proposed thresholding method with inverse HWT is conducted to extract the final feature set. Finally, the K-nearest neighbor (k-NN) algorithm is used to recognize the activity with respect to the given data. The method compares favorably with the results using other machine learning algorithms.

• Asian-Australasian Journal of Animal Sciences
• /
• v.28 no.3
• /
• pp.323-327
• /
• 2015

Wattle length, width, and area were measured to classify bilateral asymmetries in four lines of chickens. The lines were the S26 generation of White Leghorns selected for high (HAS) or low (LAS) response to sheep red blood cells and sublines in which selection had been relaxed for three generations (high antibody relaxed [HAR] and low antibody relaxed [LAR]). Antibody titers (AB) were greater for HAS than for HAR with both greater than for LAS and LAR which while different for males did not differ for females. The low antibody lines were heavier and reached sexual maturity at younger age than the high antibody lines. In general, wattle length, width, and area were greater in the low than high antibody lines. In 24 comparisons for bilaterality 18 exhibited fluctuating asymmetry and 6 exhibited directional asymmetry with 5 of the 6 being for wattle length. There was not a clear pattern for changes in degree of asymmetry when selection was relaxed for 3 generations. For females, the relative asymmetry (RA) of wattle area was larger ($p{\leq}0.05$) for HAR than for LAR and not different from the selected lines and relaxed lines. There were no differences among lines for RA of wattle length and width of females and wattle length, width, and area of males.

• Journal of the Korean Society for Precision Engineering
• /
• v.23 no.3 s.180
• /
• pp.156-162
• /
• 2006

Direct fabrication of nano patterns has been studied employing a nano-stereolithography (NSL) process. The needs of nano patterning techniques have been intensively increased for diverse applications for nano/micro-devices; micro-fluidic channels, micro-molds. and other novel micro-objects. For fabrication of high-aspect-ratio (HAR) patterns, a thick spin coating of SU-8 process is generally used in the conventional photolithography, however, additional processes such as pre- and post-baking processes and expansive precise photomasks are inevitably required. In this work, direct fabrication of HAR patterns with a high spatial resolution is tried employing two-photon polymerization in the NSL process. The precision and aspect ratio of patterns can be controlled using process parameters of laser power, exposure time, and numerical aperture of objective lens. It is also feasible to control the aspect ratio of patterns by truncation amounts of patterns, and a layer-by-layer piling up technique is attempted to achieve HAR patterns. Through the fabrication of several patterns using the NSL process, the possibility of effective patterning technique fer various N/MEMS applications has been demonstrated.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2004.08a
• /
• pp.216-222
• /
• 2004

The caster roll shells have the good thermal conductivity and the low thermal expansion and have to exhibit high enough strength and good ductility at temperature up to $600^{\circ}C$. Thermal stress in particular is very high due to the contact with the liquid aluminium. The main stresses are of thermal origin, which bring a plastic fatigue on surface. This paper will represent one survey about the investigation of the failure of roll shells for continuous casters and an analysis using the simulation of the temperature distribution and the state of stress during hot rolling. Moreover, there will be a discussion on the roll shell of Mod. HAR 5 which is developed by heat treatment process. Mod. HAR 5 has advantages of high strength, high toughness and increased thermal stress resistance while maintaining the same productivity as the conventional roll.

• Bulletin of the Korean Chemical Society
• /
• v.29 no.3
• /
• pp.656-662
• /
• 2008

Antagonists of the d -opioid receptor are effective in overcoming resistance against analgesic drugs such as morphine. To identify novel antagonists of the d -opioid receptor that display high potency and low resistance, we performed 3D-QSAR analysis using chemical feature-based pharmacophore models. Chemical features for d -opioid receptor antagonists were generated using quantitative (Catalyst/HypoGen) and qualitative (Catalyst/HipHop) approaches. For HypoGen analysis, we collected 16 peptide and 16 non-peptide antagonists as the training set. The best-fit pharmacophore hypotheses of the two antagonist models comprised identical features, including a hydrophobic aromatic (HAR), a hydrophobic (HY), and a positive ionizable (PI) function. The training set of the HipHop model was constructed with three launched opioid drugs. The best hypothesis from HipHop included four features: an HAR, an HY, a hydrogen bond donor (HBD), and a PI function. Based on these results, we confirm that HY, HAR and PI features are essential for effective antagonism of the d -opioid receptor, and determine the appropriate pharmacophore to design such antagonists.

• Journal of Internet Computing and Services
• /
• v.22 no.3
• /
• pp.9-16
• /
• 2021

This study aims to develop a human activity recognition (HAR) system as a Deep-Learning (DL) classification model, distinguishing various human activities. We solely rely on the signals from a wristband accelerometer worn by a person for the user's convenience. 3-axis sequential acceleration signal data are gathered within a predefined time-window-slice, and they are used as input to the classification system. We are particularly interested in developing a Deep-Learning model that can outperform conventional machine learning classification performance. A total of 13 activities based on the laboratory experiments' data are used for the initial performance comparison. We have improved classification performance using the Convolutional Neural Network (CNN) combined with an auto-encoder feature reduction and parameter tuning. With various publically available HAR datasets, we could also achieve significant improvement in HAR classification. Our CNN model is also compared against Recurrent-Neural-Network(RNN) with Long Short-Term Memory(LSTM) to demonstrate its superiority. Noticeably, our model could distinguish both general activities and near-identical activities such as sitting down on the chair and floor, with almost perfect classification accuracy.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.6
• /
• pp.89-95
• /
• 2023

In recent decades, human action recognition (HAR) has demonstrated potential applications in sports analysis, human-robot interaction, and large-scale signage content. In this paper, spatial temporal attention graph convolutional network (STAGCN)-based HAR system is proposed. Spatioal-temmporal features of skeleton sequences are assigned different weights by STAGCN, enabling the consideration of key joints and viewpoints. From simulation results, it has been shown that the performance of the proposed model can be improved in terms of classification accuracy in the NTU RGB+D dataset.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 1998.04a
• /
• pp.282-285
• /
• 1998

• Kyungpook Mathematical Journal
• /
• v.32 no.1
• /
• pp.45-54
• /
• 1992

• Electrical & Electronic Materials
• /
• v.18 no.4
• /
• pp.12-18
• /
• 2005