Fall Prediction of Elder Person Using CCTV Footage and Media Framework
Published in 2023 International Conference on Emerging Techniques in Computational Intelligence (ICETCI), 2023
The elderly population often requires assistance with day-to-day activities and healthcare. The development and use of efficient procedures and systems that would enable affordable clinic care and monitoring services, particularly for the senior population, are receiving more attention. The concept of “ageing in place” refers to the capability of older individuals to live in their own homes and communities comfortably, securely, and independently, without being restricted by their age, financial situation, or physical ability. Moving in with relatives, relocating to a medical facility, or entering an assisted living facility can all cause psychological distress for older adults. This stress can deteriorate their health and lower their quality of life. MediaPipe Body Landmark model employs BlazePose to infer 33 3D body landmarks or 25 upper-body landmarks from RGB video frames to provide high-quality body posture monitoring. It can be configured to focus solely on the upper body, in which case it only estimates the first 25 landmarks, but typically, it identifies landmarks for every body pose. MediaPipe Hands is an effective method for monitoring hands and fingers. A single picture may be used to infer up to 21 3D hand landmarks. Mobile devices can enable a variety of modern life applications by simultaneously detecting human gestures, face landmarks, and hand movements in real-time, such as augmented reality try-on and effects, posture control, sign language recognition, fitness and sport analysis, and posture control.MediaPipe now offers immediate, exact, distinct, yet complementary solutions for these challenging jobs. Combining them all into a real-time, conceptually cohesive edge solution requires the simultaneous inference ofmany, dependent neural networks, which makes it a particularly difficult challenge. We are proposing a low cost and efficient system by using mediapipe body landmark model for extracting the features and PCA-LSTM algorithm to predict the fall in elderly people.
Recommended citation: D. M. Nazeer, V. S. Chaitanya kolliboyina, K. K. Tiruveedula, I. s. H. Punithavathi, C. Shwetha and D. Anusha, "Fall Prediction of Elder Person Using CCTV Footage and Media Framework," 2023 International Conference on Emerging Techniques in Computational Intelligence (ICETCI), Hyderabad, India, 2023, pp. 138-144, doi: 10.1109/ICETCI58599.2023.10331422. https://ieeexplore.ieee.org/abstract/document/10331422