Abstract:
The project aims to develop the most practical design of a smart chair system that can
accurately and seamlessly monitor the seating behaviour ofthe human body. We tested this in
a routine survey because most of those we tested felt that reducing pain in an unhealthy and
uncomfortable posture and sitting could prevent or reduce chronic illness. These good results
indicate that the submitted system is possible to monitor seating behaviour, which can find
application in many areas, including healthcare, human-computer interactions, and intelligent
environments. In this smart chair system, we have incorporated many image processing
techniques into in-depth images that have been successfully captured to identify different
human postures such as sitting upright, hunchback, arm folds and more. Idle and poor sitting
posture can be detrimental to the health of workers working in offices, banks or any other
organization. Therefore, it is very practical to effectively identify the sitting posture of people
in the work environment and to warn them about the poor sitting posture. The Case Machine
Learning Framework is used to train the datasets and to identify the seating posture that the
camera uses to identify the sitting posture and image recognition to collect the posture feature.
Experiments show that the accuracy ranges from 70% to more than 75%. The HC-SR04
ultrasonic sensor with a thickness of 0.25 mm was used, which is thin enough and suitable for
application. An ultrasonic sensor detects whether a person is close to the backrest.
