Aim: IJASCI journal is a peer-reviewed open access journal that serves as a platform for the dissemination of original and cutting-edge research related to Applied sciences and computational Intelligence.
ISSN No: XXXX-XXXX
Impact Factor: x.x (2022); 5-Year Impact Factor: x.x (2024)
Published: Feb 26, 2024
DOI: 10.4018/JGIM.339238
Volume 32
HUMAN POSE ESTIMATION USING MEDIAPIPE AND OPENCV
A.V. Suvarna, Dr.C. Sivaraj
Abstract : The well-being of seniors living alone is a growing concern, particularly regarding the risk of falls and associated injuries. To mitigate these risks, there is a need for an automated mobile robot that can monitor and recognize their poses. While deep learning methods have shown progress in human pose estimation, accurately estimating poses that are infrequent or absent in training datasets remains a challenge.
In this work, we propose an estimation model that combines the capabilities of MediaPipe and OpenCV, specifically utilizing the MediaPipe’s BlazePose model to detect the 33 key points on the human body, capturing important body landmarks for pose estimation. OpenCV is then employed to further refine and enhance the estimation results. This combination enables more accurate monitoring and recognition of poses, providing valuable information for fall detection and injury prevention in seniors living alone
Published: Feb 26, 2024
DOI: 10.4018/JGIM.339238
Volume 32
HUMAN POSE ESTIMATION USING MEDIAPIPE AND OPENCV
A.V. Suvarna, Dr.C. Sivaraj
Abstract : The well-being of seniors living alone is a growing concern, particularly regarding the risk of falls and associated injuries. To mitigate these risks, there is a need for an automated mobile robot that can monitor and recognize their poses. While deep learning methods have shown progress in human pose estimation, accurately estimating poses that are infrequent or absent in training datasets remains a challenge. In this work, we propose an estimation model that combines the capabilities of MediaPipe and OpenCV, specifically utilizing the MediaPipe’s BlazePose model to detect the 33 key points on the human body, capturing important body landmarks for pose estimation. OpenCV is then employed to further refine and enhance the estimation results. This combination enables more accurate monitoring and recognition of poses, providing valuable information for fall detection and injury prevention in seniors living alone
Published: Feb 26, 2024
DOI: 10.4018/JGIM.339238
Volume 32
HUMAN POSE ESTIMATION USING MEDIAPIPE AND OPENCV
A.V. Suvarna, Dr.C. Sivaraj
Abstract : The well-being of seniors living alone is a growing concern, particularly regarding the risk of falls and associated injuries. To mitigate these risks, there is a need for an automated mobile robot that can monitor and recognize their poses. While deep learning methods have shown progress in human pose estimation, accurately estimating poses that are infrequent or absent in training datasets remains a challenge. In this work, we propose an estimation model that combines the capabilities of MediaPipe and OpenCV, specifically utilizing the MediaPipe’s BlazePose model to detect the 33 key points on the human body, capturing important body landmarks for pose estimation. OpenCV is then employed to further refine and enhance the estimation results. This combination enables more accurate monitoring and recognition of poses, providing valuable information for fall detection and injury prevention in seniors living alone
Published: Feb 26, 2024
DOI: 10.4018/JGIM.339238
Volume 32
HUMAN POSE ESTIMATION USING MEDIAPIPE AND OPENCV
A.V. Suvarna, Dr.C. Sivaraj
Abstract : The well-being of seniors living alone is a growing concern, particularly regarding the risk of falls and associated injuries. To mitigate these risks, there is a need for an automated mobile robot that can monitor and recognize their poses. While deep learning methods have shown progress in human pose estimation, accurately estimating poses that are infrequent or absent in training datasets remains a challenge. In this work, we propose an estimation model that combines the capabilities of MediaPipe and OpenCV, specifically utilizing the MediaPipe’s BlazePose model to detect the 33 key points on the human body, capturing important body landmarks for pose estimation. OpenCV is then employed to further refine and enhance the estimation results. This combination enables more accurate monitoring and recognition of poses, providing valuable information for fall detection and injury prevention in seniors living alone