Low-Cost Non-Wearable Fall Detection System Implemented on a Single Board Computer for People in Need of Care

Abstract

This work aims at proposing an affordable, non-wearable system to detect falls of people in need of care. The proposal uses artificial vision based on deep learning techniques implemented on a Raspberry Pi4 4GB RAM with a High-Definition IR-CUT camera. The CNN architecture classifies detected people into five classes: fallen, crouching, sitting, standing, and lying down. When a fall is detected, the system sends an alert notification to mobile devices through the Telegram instant messaging platform. The system was evaluated considering real daily indoor activities under different conditions: outfit, lightning, and distance from camera. Results show a good trade-off between performance and cost of the system. Obtained performance metrics are: precision of 96.4%, specificity of 96.6%, accuracy of 94.8%, and sensitivity of 93.1%. Regarding privacy concerns, even though this system uses a camera, the video is not recorded or monitored by anyone, and pictures are only sent in case of fall detection. This work can contribute to reducing the fatal consequences of falls in people in need of care by providing them with prompt attention. Such a low-cost solution would be desirable, particularly in developing countries with limited or no medical alert systems and few resources.

Keywords: CNN; artificial vision; deep-learning; fall-detection; non-wearable; people in need of care; single board computer.

Figure 1

Methodology.

Figure 2

SSD Mobilenetv2 architecture.

Figure 3

Total loss curves.

Figure 4

Embedded application block diagram.

Figure 5

System flowchart.

Figure 6

Classes detected: Fallen person (“Persona caída”), Crouching person (“Persona agachada”), Sitting person (“Persona sentada”), Standing person (“Persona parada”), Person lying down (“Persona acostada”).

Figure 7

Different poses of a “Fallen person”.

Figure 8

Prediction results in living room considering contextual information.

Figure 9

Prediction results in bedroom considering contextual information.

Figure 10

Test design under different outfit, lightning, and distance conditions.

Figure 11

System performance with different outfits.

Figure 12

System performance under different lightning levels.

Figure 13

System performance at different distances from the camera.

Figure 14

Confusion matrix of tested scenarios.

Figure 15

ROC curve of the proposed system.