• Journal of KIISE:Software and Applications
• /
• v.41 no.8
• /
• pp.586-597
• /
• 2014

The next generation mobile computing technology is recently attracting attention that smartphone and wearable device imbedded with various sensors are being deployed in the world. Existing activity and posture recognition research can be divided into two different ways considering feature of one's movement. While activity recognition focuses on catching distinct pattern according to continuous movement, posture recognition focuses on sudden change of posture and body orientation. There is a lack of research constructing a system mixing two separate patterns which could be applied in real world. In this paper, we propose a method to use both smartphone and wearable device to recognize activity and posture in the same time. To use smartphone and wearable sensor data together, we designed a pre-processing method and constructed recognition model mixing signal vector magnitude and orientation pattern features of vertical and horizontal. We considered cycling, fast/slow walking and running activities, and postures such as standing, sitting, and laying down. We confirmed the performance and validity by experiment, and proved the feasibility in real world.

• Journal of Sensor Science and Technology
• /
• v.16 no.6
• /
• pp.467-474
• /
• 2007

The real-time monitoring about the activity of the human provides useful information about the activity quantity and ability. The present study implemented a small-size and low-power acceleration monitoring system for convenient monitoring of activity quantity and recognition of emergent situations such as falling during daily life. For the wireless transmission of acceleration sensor signal, we developed a wireless transmission system based on a wireless sensor network. In addition, we developed a program for storing and monitoring wirelessly transmitted signals on PC in real-time. The performance of the implemented system was evaluated by assessing the output characteristic of the system according to the change of posture, and parameters and acontext recognition algorithm were developed in order to monitor activity volume during daily life and to recognize emergent situations such as falling. In particular, recognition error in the sudden change of acceleration was minimized by the application of a falling correction algorithm

• The KIPS Transactions:PartB
• /
• v.18B no.5
• /
• pp.305-314
• /
• 2011

Research into the visual perception of human emotion to recognize an intention has traditionally focused on emotions of facial expression. Recently researchers have turned to the more challenging field of emotional expressions through body posture or activity. Proposed work approaches recognition of basic emotional categories from body postures using neural model applied visual perception of neurophysiology. In keeping with information processing models of the visual cortex, this work constructs a biologically plausible hierarchy of neural detectors, which can discriminate 6 basic emotional states from static views of associated body postures of activity. The proposed model, which is tolerant to parameter variations, presents its possibility by evaluating against human test subjects on a set of body postures of activities.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.5
• /
• pp.177-183
• /
• 2022

Recently, interesting of home training is getting bigger due to COVID-19. Accordingly, research on applying HAR(human activity recognition) technology to home training has been conducted. However, existing paper of HAR proposed static activity instead of dynamic activity. In this paper, the deep learning model where dynamic exercise posture can be analyzed and the accuracy of the user's exercise posture can be shown is proposed. Fitness images of AI-hub are analyzed by blaze pose. The experiment is compared with three types of deep learning model: RNN(recurrent neural network), LSTM(long short-term memory), CNN(convolution neural network). In simulation results, it was shown that the f1-score of RNN, LSTM and CNN is 0.49, 0.87 and 0.98, respectively. It was confirmed that CNN is more suitable for human activity recognition than other models from simulation results. More exercise postures can be analyzed using a variety learning data.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.877-885
• /
• 2022

Vision-based activity recognition has been widely attempted at construction sites to estimate productivity and enhance workers' health and safety. Previous studies have focused on extracting an individual worker's postural information from sequential image frames for activity recognition. However, various trades of workers perform different tasks with similar postural patterns, which degrades the performance of activity recognition based on postural information. To this end, this research exploited a concept of human-object interaction, the interaction between a worker and their surrounding objects, considering the fact that trade workers interact with a specific object (e.g., working tools or construction materials) relevant to their trades. This research developed an approach to understand the context from sequential image frames based on four features: posture, object, spatial features, and temporal feature. Both posture and object features were used to analyze the interaction between the worker and the target object, and the other two features were used to detect movements from the entire region of image frames in both temporal and spatial domains. The developed approach used convolutional neural networks (CNN) for feature extractors and activity classifiers and long short-term memory (LSTM) was also used as an activity classifier. The developed approach provided an average accuracy of 85.96% for classifying 12 target construction tasks performed by two trades of workers, which was higher than two benchmark models. This experimental result indicated that integrating a concept of the human-object interaction offers great benefits in activity recognition when various trade workers coexist in a scene.

• Journal of the Korean Society of Industry Convergence
• /
• v.26 no.1
• /
• pp.113-119
• /
• 2023

In this paper, we propose an optimization of a pose estimation deep learning model for recognition of abnormal behavior in unmanned stores using radio frequencies. The radio frequency use millimeter wave in the 30 GHz to 300 GHz band. Due to the short wavelength and strong straightness, it is a frequency with less grayness and less interference due to radio absorption on the object. A millimeter wave radar is used to solve the problem of personal information infringement that may occur in conventional CCTV image-based pose estimation. Deep learning-based pose estimation models generally use convolution neural networks. The convolution neural network is a combination of convolution layers and pooling layers of different types, and there are many cases of convolution filter size, number, and convolution operations, and more cases of combining components. Therefore, it is difficult to find the structure and components of the optimal posture estimation model for input data. Compared with conventional millimeter wave-based posture estimation studies, it is possible to explore the structure and components of the optimal posture estimation model for input data using genetic algorithms, and the performance of optimizing the proposed posture estimation model is excellent. Data are collected for actual unmanned stores, and point cloud data and three-dimensional keypoint information of Kinect Azure are collected using millimeter wave radar for collapse and property damage occurring in unmanned stores. As a result of the experiment, it was confirmed that the error was moored compared to the conventional posture estimation model.

• Journal of Korea Multimedia Society
• /
• v.14 no.8
• /
• pp.992-999
• /
• 2011

This paper proposed a human body posture recognition program based on haar-like feature and hand activity detection. Its distinguishing features are the combination of face detection and motion detection. Firstly, the program uses the haar-like feature face detection to receive the location of human face. The haar-like feature is provided with the advantages of speed. It means the less amount of calculation the haar-like feature can exclude a large number of interference, and it can discriminate human face more accurately, and achieve the face position. Then the program uses the frame subtraction to achieve the position of human body motion. This method is provided with good performance of the motion detection. Afterwards, the program recognises the human body motion by calculating the relationship of the face position with the position of human body motion contour. By the test, we know that the recognition rate of this algorithm is more than 92%. The results show that, this algorithm can achieve the result quickly, and guarantee the exactitude of the result.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.286-299
• /
• 2022

Human activity recognition in real time is a challenging task. Recently, a plethora of studies has been proposed using deep learning architectures. The implementation of these architectures requires the high computing power of the machine and a massive database. However, handcrafted features-based machine learning models need less computing power and very accurate where features are effectively extracted. In this study, we propose a handcrafted model based on three-dimensional sequential skeleton data. The human body skeleton movement over a frame is computed through joint positions in a frame. The joints of these skeletal frames are projected into two-dimensional space, forming a "movement polygon." These polygons are further transformed into a one-dimensional space by computing amplitudes at different angles from the centroid of polygons. The feature vector is formed by the sampling of these amplitudes at different angles. The performance of the algorithm is evaluated using a support vector machine on four public datasets: MSR Action3D, Berkeley MHAD, TST Fall Detection, and NTU-RGB+D, and the highest accuracies achieved on these datasets are 94.13%, 93.34%, 95.7%, and 86.8%, respectively. These accuracies are compared with similar state-of-the-art and show superior performance.

• Smart Media Journal
• /
• v.1 no.2
• /
• pp.12-22
• /
• 2012

Today we are witnessing an increasingly widespread use of cameras in our lives for video surveillance, robot vision, and mobile phones. This has led to a renewed interest in computer vision in general and an on-going boom in human activity recognition in particular. Although not particularly fancy per se, human gait is inarguably the most common and frequent action. Early on this decade there has been a passing interest in human gait recognition, but it soon declined before we came up with a systematic analysis and understanding of walking motion. This paper presents a set of DBN-based models for the analysis of human gait in sequence of increasing complexity and modeling power. The discussion centers around HMM-based statistical methods capable of modeling the variability and incompleteness of input video signals. Finally a novel idea of extending the discrete state Markov chain with a continuous density function is proposed in order to better characterize the gait direction. The proposed modeling framework allows us to recognize pedestrian up to 91.67% and to elegantly decode out two independent gait components of direction and posture through a sequence of experiments.

• Journal of Digital Convergence
• /
• v.18 no.9
• /
• pp.191-198
• /
• 2020

In this study, in order to propose an improvement plan for fire simulation training to improve fire response capability, fire simulation training is conducted for three business facilities, and changes according to whether training materials are used and whether the response time for each response stage is reflected in the training evaluation. The response posture and response ability of the training participants were observed and analyzed. As a result of the analysis, it was analyzed that most of the training participants improved their participation in training, response posture, and response ability when the use of training textbooks and response time for each response stage were reflected in the training evaluation. In the event of a fire simulation training, a number of training materials that can similarly implement the fire situation are used to improve and maintain the fire response capabilities (fire notification and fire report, initial extinguishing, and evacuation) of the training participants, and the target time for each response step. The result was that it can be used as a useful index for improving fire response capability and improving fire simulation training in the future and feedback only when quantitative training evaluation is conducted based on this setting.