• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.35-36
• /
• 2022

본 논문은 딥러닝 사람 자세 추정 모델이 사람의 관절 키포인트를 예측하는데 관절의 2차원 면적에 의해 키포인트별 𝜎, 즉, 표준 편차를 가지는 가우시안 커널(Gaussian Kernel)을 예측하는 방법을 제안한다. 각 관절 키포인트에 대해 다른 𝜎를 가지는 정답 히트맵(Ground Truth Heatmap)과 제안한 Gaussian Mixture Block를 모델에 추가해서 관절의 크기를 맞는 히트맵을 예측한다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.6
• /
• pp.75-83
• /
• 2007

A model based approach of tracking the limbs of a walking human subject is proposed in this paper. The tracking process begins by building a data base composed of conditional probabilities of motions between the limbs of a walking subject. With a suitable amount of video footage from various human subjects included in the database, a probabilistic model characterizing the relationships between motions of limbs is developed. The motion tracking of a test subject begins with identifying and tracking limbs from the surveillance video image using the edge and silhouette detection methods. When occlusion occurs in any of the limbs being tracked, the approach uses the probabilistic motion model in conjunction with the minimum cost based edge and silhouette tracking model to determine the motion of the limb occluded in the image. The method has shown promising results of tracking occluded limbs in the validation tests.

• Journal of Digital Convergence
• /
• v.11 no.12
• /
• pp.295-302
• /
• 2013

Many reported methods assume that the people in an image or an image sequence have been identified and localization. People detection is one of very important variable to affect for the system's performance as the basis technology about the detection of other objects and interacting with people and computers, motion recognition. In this paper, we present an efficient linear discriminant for multi-view people detection. Our approaches are based on linear discriminant. We define training data with fisher Linear discriminant to efficient learning method. People detection is considerably difficult because it will be influenced by poses of people and changes in illumination. This idea can solve the multi-view scale and people detection problem quickly and efficiently, which fits for detecting people automatically. In this paper, we extract people using fisher linear discriminant that is hierarchical models invariant pose and background. We estimation the pose in detected people. The purpose of this paper is to classify people and non-people using fisher linear discriminant.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.89-90
• /
• 2021

본 논문에서는 3D(dimensional) 스켈레톤을 이용하여 멀티 뷰 RGB-D 카메라를 캘리브레이션 하는 새로운 기법을 제안하고자 한다. 멀티 뷰 카메라를 캘리브레이션 하기 위해서는 일관성 있는 특징점이 필요하다. 우리는 다시점 카메라를 캘리브레이션 하기 위한 특징점으로 사람의 스켈레톤을 사용한다. 사람의 스켈레톤은 최신의 자세 추정(pose estimation) 알고리즘들을 이용하여 쉽게 구할 수 있게 되었다. 우리는 자세 추정 알고리즘을 통해서 획득된 3D 스켈레톤의 관절 좌표를 특징점으로 사용하는 RGB-D 기반의 캘리브레이션 알고리즘을 제안한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.12
• /
• pp.1528-1534
• /
• 2019

Recently, due to the popularization of 3D depth cameras, new researches and opportunities have been made in research conducted on RGB images, but estimation of human hand pose is still classified as one of the difficult topics. In this paper, we propose a robust estimation method of human hand pose from various input 3D depth images using a learning algorithm. The proposed approach first generates a skeleton-based hand model and then aligns the generated hand model with three-dimensional point cloud data. Then, using a random forest-based learning algorithm, the hand pose is strongly estimated from the aligned hand model. Experimental results in this paper show that the proposed hierarchical approach makes robust and fast estimation of human hand posture from input depth images captured in various indoor and outdoor environments.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.9
• /
• pp.17-25
• /
• 2023

In this paper, we propose a model that can perform human pose estimation through a MobileViT-based model with fewer parameters and faster estimation. The based model demonstrates lightweight performance through a structure that combines features of convolutional neural networks with features of Vision Transformer. Transformer, which is a major mechanism in this study, has become more influential as its based models perform better than convolutional neural network-based models in the field of computer vision. Similarly, in the field of human pose estimation, Vision Transformer-based ViTPose maintains the best performance in all human pose estimation benchmarks such as COCO, OCHuman, and MPII. However, because Vision Transformer has a heavy model structure with a large number of parameters and requires a relatively large amount of computation, it costs users a lot to train the model. Accordingly, the based model overcame the insufficient Inductive Bias calculation problem, which requires a large amount of computation by Vision Transformer, with Local Representation through a convolutional neural network structure. Finally, the proposed model obtained a mean average precision of 0.694 on the MS COCO benchmark with 3.28 GFLOPs and 9.72 million parameters, which are 1/5 and 1/9 the number compared to ViTPose, respectively.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.1048-1051
• /
• 2020

본 논문에서는 사람의 손에 관한 개별적으로 분리되어 진행되고 있는 손 위치 추정, 손 자세 추정, 손 동작 인식 작업을 통합하는 Faster-RCNN기반의 프레임워크를 제안하였다. 제안된 프레임워크에서는 RGB 동영상을 입력으로 하여, 먼저 손 위치에 대한 박스를 생성하고, 생성된 박스 정보를 기반으로 손 자세와 동작을 인식하도록 한다. 손 위치, 손 자세, 손 동작에 대한 정답을 동시에 모두 가지는 데이터셋이 존재하지 않기 때문에 Egohands, FPHA 데이터를 동시에 효과적으로 사용하는 방안을 제안하였으며 제안된 프레임워크를 FPHA데이터에 평가하였다., 손 위치 추정 정확도는 mAP 90.3을 기록했고, 손 동작 인식은 FPHA의 정답을 사용한 정확도에 근접한 70.6%를 기록하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.189-196
• /
• 2024

Frequent false positives alarm from the Intelligent Selective Control System have raised significant concerns. These persistent issues have led to declines in operational efficiency and market credibility among agents. Developing a new model or replacing the existing one to mitigate false positives alarm entails substantial opportunity costs; hence, improving the quality of the training dataset is pragmatic. However, smaller organizations face challenges with inadequate capabilities in dataset collection and refinement. This paper proposes an automatic human pose data collection system centered around a human pose estimation model, utilizing camera-based sensor fusion techniques and edge devices. The system facilitates the direct collection and real-time processing of field data at the network periphery, distributing the computational load that typically centralizes. Additionally, by directly labeling field data, it aids in constructing new training datasets.

• Journal of the Society of Disaster Information
• /
• v.19 no.4
• /
• pp.897-907
• /
• 2023

Purpose: In this study, We aim to estimate a untrained person's three postures using a 2D CNN model which is trained with minimal FFT data collected by a 24GHz FMCW radar. Method: In an indoor space, we collected FFT data for three distinct postures (standing, sitting, and lying) from three different individuals. To apply this data to a 2D CNN model, we first converted the collected data into 2D images. These images were then trained using the 2D CNN model to recognize the distinct features of each posture. Following the training, we evaluated the model's accuracy in differentiating the posture features across various individuals. Result: According to the experimental results, the average accuracy of the proposed scheme for the three postures was shown to be a 89.99% and it outperforms the conventional 1D CNN and the SVM schemes. Conclusion: In this study, we aim to estimate any person's three postures using a 2D CNN model and a 24GHz FMCW radar for disastrous situations in indoor. it is shown that the different posture of any persons can be accurately estimated even though his or her data is not used for training the AI model.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.369-372
• /
• 2021

These days, people spend more time sitting at a desk for studies or work. Also, because people continue to use computers, smartphones, and tablet PCs often during break times, their posture is getting worse. Maintaining a position of bad posture for an extended period of time causes problems with the musculoskeletal system related to the neck, shoulders, and spine. Additionally, problems such as physical fatigue and posture deformation are predicted to expand to a wide range of age groups. Therefore, the core function of the system we are developing is to ensure correct sitting posture and to receive alert notifications via the created mobile application. To create the system, a flex sensor, pressure sensor, and tilt sensor are attached to a chair and utilized. The flex sensor detects and compares the amount of bending in the chair's posture and transmits this value to an Arduino Uno R3 board. Additionally, information such as body balance and incline angle are collected to determine whether or not the current sitting posture is correct. When the posture is incorrect, a notification is sent through the mobile application to indicate to the user and the monitoring app that their posture is not correct. The system proposed in this study is expected to be of great help in future posture-related research.