An Elderly Person Protection System Based on Dynamic Image Processing

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K23333
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: Ritsumeikan University
• Principal Investigator: Kong Xiangbo 立命館大学, 理工学部, 助教 (20880404)
• Project Period (FY): 2020-09-11 – 2023-03-31
• Keywords: 画像処理 / 高齢者の見守り

Outline of Final Research Achievements

Falls among the elderly can pose a serious risk to their health and well-being if not promptly detected. To address this issue, we proposed a protection system for elderly people, with a specific focus on fall detection in this study. Traditional approaches rely on depth images to detect joint coordinates in 3D space and recognize posture. However, depth-based systems are costly. Alternatively, color-based monitoring systems raise privacy concerns. To mitigate these challenges, we introduced a technique for estimating depth images from color images, which helps to address privacy concerns in the fall detection system. Additionally, we conducted research on implementing image processing algorithms on edge devices to achieve high speed and low power consumption.

Free Research Field

画像処理

Academic Significance and Societal Importance of the Research Achievements

一人暮らしの高齢者は、筋力の低下や、歩行障害、視力の衰え、心疾患、脳血管疾患などの原因によって転倒し、発見が遅れると、命の危険もある。画像処理型の見守りシステムは、ウェアラブルデバイスを着用する必要がない、認識率が高いなどの利点がある。しかし、画像処理には、プライバシ侵害の恐れがある。本研究では、ぼやけて見える深度画像に基づく高齢者の見守りシステムを提案し、従来研究のプライバシ問題を一定程度解決した。また、本研究で提案した深度推定技術は、屋内・屋外の転倒検出だけでなく、他の研究分野でも期待できる。