• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.3
• /
• pp.155-160
• /
• 2017

Recently, Deep learning has been used successfully to solve many recognition problems. It has many advantages over existing machine learning methods that extract feature points through hand-crafting. Deep neural networks for human activity recognition split video data into frame images, and then classify activities by analysing the connectivity of frame images according to the time. But it is difficult to apply to actual problems which has small-scale activity classes. Because this situations has a problem of overfitting and insufficient training data. In this paper, we defined 5 type of small-scale human activities, and classified them. We construct video database using 700 video clips, and obtained a classifying accuracy of 74.00%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.11
• /
• pp.2853-2873
• /
• 2013

Over the last few years, one of the most common purposes of smart homes is to provide human centric services in the domain of u-healthcare by analyzing inhabitants' daily living. Currently, the major challenges in activity recognition include the reliability of prediction of each classifier as they differ according to smart homes characteristics. Smart homes indicate variation in terms of performed activities, deployed sensors, environment settings, and inhabitants' characteristics. It is not possible that one classifier always performs better than all the other classifiers for every possible situation. This observation has motivated towards combining multiple classifiers to take advantage of their complementary performance for high accuracy. Therefore, in this paper, a method for activity recognition is proposed by optimizing the output of multiple classifiers with Genetic Algorithm (GA). Our proposed method combines the measurement level output of different classifiers for each activity class to make up the ensemble. For the evaluation of the proposed method, experiments are performed on three real datasets from CASAS smart home. The results show that our method systematically outperforms single classifier and traditional multiclass models. The significant improvement is achieved from 0.82 to 0.90 in the F-measures of recognized activities as compare to existing methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.6
• /
• pp.2070-2086
• /
• 2014

As a kind of personal lifelog data, activity data have been considered as one of the most compelling information to understand the user's habits and to calibrate diagnoses. In this paper, we proposed a robust algorithm to sampling rates for human activity recognition, which identifies a user's activity using accelerations from a triaxial accelerometer in a smartphone. Although a high sampling rate is required for high accuracy, it is not desirable for actual smartphone usage, battery consumption, or storage occupancy. Activity recognitions with well-known algorithms, including MLP, C4.5, or SVM, suffer from a loss of accuracy when a sampling rate of accelerometers decreases. Thus, we start from particle swarm optimization (PSO), which has relatively better tolerance to declines in sampling rates, and we propose PSO with an adaptive boundary correction (ABC) approach. PSO with ABC is tolerant of various sampling rate in that it identifies all data by adjusting the classification boundaries of each activity. The experimental results show that PSO with ABC has better tolerance to changes of sampling rates of an accelerometer than PSO without ABC and other methods. In particular, PSO with ABC is 6%, 25%, and 35% better than PSO without ABC for sitting, standing, and walking, respectively, at a sampling period of 32 seconds. PSO with ABC is the only algorithm that guarantees at least 80% accuracy for every activity at a sampling period of smaller than or equal to 8 seconds.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.443-452
• /
• 2020

Worker's awkward postures and unreasonable physical load can be corrected by monitoring construction activities, thereby increasing the safety and productivity of construction workers and projects. However, manual identification is time-consuming and contains high human variance. In this regard, an automated activity recognition system based on inertial measurement unit can help in rapidly and precisely collecting motion data. With the acceleration data, the machine learning algorithm will be used to train classifiers for automatically categorizing activities. However, input acceleration data are extracted either from designed experiments or simple construction work in previous studies. Thus, collected data series are discontinuous and activity categories are insufficient for real construction circumstances. This study aims to collect acceleration data during long-term continuous work in a construction project and validate the feasibility of activity recognition algorithm with the continuous motion data. The data collection covers two different workers performing formwork at the same site. An accelerator, as well as portable camera, is attached to the worker during the entire working session for simultaneously recording motion data and working activity. The supervised machine learning-based models are trained to classify activity in hierarchical levels, which reaches a 96.9% testing accuracy of recognizing rest and work and 85.6% testing accuracy of identifying stationary, traveling, and rebar installation actions.

• Journal of Korea Multimedia Society
• /
• v.24 no.3
• /
• pp.416-422
• /
• 2021

Due to the large differences in human activity within classes, the large similarity between classes, and the problems of visual angle and occlusion, it is difficult to extract features manually, and the detection rate of human behavior is low. In order to better solve these problems, an improved Faster R-CNN-based detection algorithm is proposed in this paper. It achieves multi-object recognition and localization through a second-order detection network, and replaces the original feature extraction module with Dense-Net, which can fuse multi-level feature information, increase network depth and avoid disappearance of network gradients. Meanwhile, the proposal merging strategy is improved with Soft-NMS, where an attenuation function is designed to replace the conventional NMS algorithm, thereby avoiding missed detection of adjacent or overlapping objects, and enhancing the network detection accuracy under multiple objects. During the experiment, the improved Faster R-CNN method in this article has 84.7% target detection result, which is improved compared to other methods, which proves that the target recognition method has significant advantages and potential.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.6
• /
• pp.2377-2397
• /
• 2020

Human activity recognition is widely used in smart homes, health care and indoor monitor. Traditional approaches all need hardware installation or wearable sensors, which incurs additional costs and imposes many restrictions on usage. Therefore, this paper presents a novel device-free activities recognition system based on the advanced wireless technologies. The fine-grained information channel state information (CSI) in the wireless channel is employed as the indicator of human activities. To improve accuracy, both amplitude and phase information of CSI are extracted and shaped into feature vectors for activities recognition. In addition, we discuss the classification accuracy of different features and select the most stable features for feature matrix. Our experimental evaluation in two laboratories of different size demonstrates that the proposed scheme can achieve an average accuracy over 95% and 90% in different scenarios.

• Journal of Creative Information Culture
• /
• v.7 no.4
• /
• pp.235-242
• /
• 2021

Recently, many human activity recognition(HAR) researches using smartphone sensor data have been studied. HAR can be utilized in various fields, such as life pattern analysis, exercise measurement, and dangerous situation detection. However researches have been focused on recognition of basic human behaviors or efficient battery use. In this paper, exercising activities performed indoors and outdoors were defined and recognized. Data collection and pre-processing is performed to recognize the defined activities by SVM, random forest and gradient boosting model. In addition, the recognition result is determined based on voting class approach for accuracy and stable performance. As a result, the proposed activities were recognized with high accuracy and in particular, similar types of indoor and outdoor exercising activities were correctly classified.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.360-363
• /
• 2018

Human activity is influenced by various factors, from individual physical features such as vertebral flexion and pelvic distortion to feelings such as joy, anger, and sadness. However, the nature of these behaviors changes over time, and behavioral characteristics do not change much in the short term. The activity data of a person has a time series characteristic that changes with time and a certain regularity for each action. In this study, we applied LSTM, a kind of cyclic neural network to deal with time - series characteristics, to the technique of recognizing activity type and improved recognition rate of activity type by measuring time and parameter optimization of components of LSTM model.

• Animal cells and systems
• /
• v.8 no.2
• /
• pp.125-133
• /
• 2004

We have extended our previous work that cross-linking CD4 molecules using specific MAb induced antigen nonspecific, MHC unrestricted killing of virally infected target cells by CD$4^+$We have extended our previous work that cross-linking CD$4^+$ molecules using specific MAb induced antigen nonspecific, MHC unrestricted killing of virally infected target cells by CD$4^+$ T cells. The killing activity of antibody activated CD$4^+$T cells was completely blocked by herbimycin A, a protein tyrosine kinase (PTK) inhibitor, but not by bisindolylamaleimide, a protein kinase C (PKC) inhibitor. Herbimycin A treated human or bovine peripheral blood CD$4^+$T cells lacked PTK activity and failed to kill virally infected target cells even after cross-linking of CD4 molecules. The CD$4^+$cross-linking failed to induce effector cell proliferation or the transcription of TNF${\beta}$ Upregulation of TNF${\beta}$ was induced by incubating the antibody activated effector cells with BHV-1 infected D17 target cells for 10 h. Anti-TNF${\beta}$ antibody partially abolished (13-44%) the direct effector cell-mediated antiviral cytotoxicity. However, this antibody neutralized 70 to 100% of antiviral activity of effector and target cell culture supernatants against BHV-1 infected D17 cells. The inhibition level of the antiviral activity by the antibody was dependent on the effector and target cell ratio. These results support the hypothesis that increased p$56^ICK enzyme activity in effector cells transduces a signal critical for effector cell recognition of viral glycoproteins expressed on the target cells. Following target cell recognition, lytic cytokines known to participate in target cell killing were produced. A better understanding of the killing activity displayed by CD$4^+$T lymphocytes following surface receptor cross-linking will provide insight into the mechanisms of cytotoxic activity directed toward virally-infected cells.T cells. The killing activity of antibody activated CD$4^+$T cells was completely blocked by herbimycin A, a protein tyrosine kinase (PTK) inhibitor, but not by bisindolylamaleimide, a protein kinase C (PKC) inhibitor. Herbimycin A treated human or bovine peripheral blood CD4T cells lacked PTK activity and failed to kill virally infected target cells even after cross-linking of CD4molecules. The CD4 cross-linking failed to induce effector cell proliferation or the transcription of TNF$\beta$. Upregulation of TNF$\beta$ was induced by incubating the antibody activated effector cells with BHV-1 infected D17 target cells for 10 h. Anti-TNF$\beta$ antibody partially abolished (13-44%) the direct effector cell-mediated antiviral cytotoxicity. However, this antibody neutralized 70 to 100% of antiviral activity of effector and target cell culture supernatants against BHV-1 infected D17 cells. The inhibition level of the antiviral activity by the antibody was dependent on the effector and target cell ratio. These results support the hypothesis that increased $56^ICK enzyme activity in effector cells transduces a signal critical for effector cell recognition of viral glycoproteins expressed on the target cells. Following target cell recognition, lytic cytokines known to participate in target cell killing were produced. A better understanding of the killing activity displayed by CD$4^+$T lymphocytes following surface receptor cross-linking will provide insight into the mechanisms of cytotoxic activity directed toward virally-infected cells.

• 한국HCI학회:학술대회논문집
• /
• 2007.02a
• /
• pp.750-754
• /
• 2007

본 논문은 실내의 로비나 복도에 설치된 방범 카메라로부터 얻어진 일련의 영상으로부터 '걷기', '뛰기', '앉기', '일어서기', '넘어짐'의 비교적 짧은 시간에 일어나는 인간 행동들을 실시간으로 인식하는 시스템의 구현에 관해 다룬다. 먼저 입력으로 받은 영상을 계층적 색인 구조를 갖는 다중 가우시안 기반의 배경 모델을 이용하여 윤곽을 추출하고 객체를 인식하여 시간차에 의한 가중치로 누적하여 시간 템플릿을 만든다. 만들어진 시간 템플릿으로부터 특징을 추출하여 신경망 모델에 적용하여 5가지 인간행동을 구분한다. 구현된 시스템으로 인간행동 인식 실험을 수행하였는데, 실험 참가자들의 행동 방식이 약간씩 달랐음에도 불구하고 높은 인식률을 보여주었다.