15.1.1. Introduction of posture

Posture is commonly defined as a position that an individual adopts while performing any functional activities such as walking, sitting, standing, sleeping or at rest. The term postural hygiene is the set of rules that aims to maintain the correct body position, either in static or in dynamic position around the head–spine–pelvis axis with total weight distributed equally on both feet (Conroy et al. 2022). Poor posture resembles the opposite sense, for example, unequal distribution of weight, the head–spine–pelvis axis is not straight or is tilted and is not aligned (Rybski and Juckett 2024).

The McKenzie method, developed by physiotherapist Robin McKenzie in the 1960s, is a comprehensive approach for diagnosing and treating musculoskeletal conditions, particularly those related to the spine. One of the core concepts of the McKenzie method is posture and its significant impact on spinal health. McKenzie emphasized that poor posture, often due to prolonged sitting, improper alignment or repetitive strain, can lead to spinal dysfunction, pain and discomfort. He introduced the idea of centralization, where movement of the spine or extremities is used to reduce pain and promote healing by shifting discomfort from peripheral areas (such as the limbs) back towards the center of the body, where it can be more effectively managed. The McKenzie approach advocates self-treatment, empowering patients to learn correct postural alignment and perform exercises designed to reduce pain, improve mobility and strengthen the muscles surrounding the spine. By focusing on posture and spinal alignment, the McKenzie method aims to not only relieve pain but also to prevent recurrence and improve overall functionality.

The McKenzie method classifies musculoskeletal disorders primarily into three syndromes: postural syndrome, dysfunction syndrome and derangement syndrome. These classifications are based on the patterns of symptoms, the mechanisms causing them and the appropriate treatment approaches.

15.1.2. Introduction to posture detectors

A posture detector is a device that detects human posture either in static or dynamic movement. This machine is composed of several sensor components, for example, wearable cameras, joint position sensors, angle sensors and tilt sensors.

These sensors can be placed on various surface landmarks of the human body. It can be noted if there is movement or an abnormal position by the computer reading against a software created grid. The software within the computer detects the stable or moving sensor and forms a picture of only sensors. These sensors can be either left individual to assess each joint and its motion or can be connected with lines to assess the coordinated movement pattern of the human body.

Below is one such image to show the representation of the behind-the-lab posture assessment and the in-software sensor attachment with lines (Piñero-Fuentes et al. 2021).

There have been many devices made to assess change in posture and give feedback to the wearer.

These devices usually have angle sensors which detect the sense of change in angle from set neutral position to either increasing or decreasing. The use of machine learning (ML) and deep learning techniques has become vital (Piñero-Fuentes et al. 2021).

The recently developed deep learning-based processing techniques process the position of various joints with convolutional neural networks (CNN) to approximate the posture and detect if any changes are sensed. This work is feasible for skeletal detection and tracking systems, aiming for proper posture guidance. Thus, posture detectors are beneficial in avoiding or preventing the musculoskeletal disorders caused by poor posture (Piñero-Fuentes et al. 2021).

Posture detectors are beneficial for various applications: identifying musculoskeletal disorders caused by prolonged static loading on intervertebral discs; aiding muscle rehabilitation for patients with motor dysfunction; and promoting workplace safety by preventing injury through posture monitoring in factories and offices. They are also useful for the treatment, prevention and monitoring of spinal disorders, such as scoliosis, where irregular posture leads to misalignment in the shoulders, waist and hips, and kyphosis, which causes a rounded or stooped back.

These detectors use wearable technology to collect movement and posture data, as well as vision cameras to monitor multiple individuals simultaneously. However, sensor-based posture detectors are contraindicated for people with metal implants, open wounds, ulcers or burns, as these can distort sensor readings and affect data accuracy (Nascimento et al. 2020; Piñero-Fuentes et al. 2021; Huang et al. 2023).

15.1.3. Working principle

• Posture detection system: a real-time pose estimation software called TRT_Pose was used by the authors. It was designed to run on NVIDIA devices, whose main purpose was to obtain the performance matrix and determine the best platform of work for ergonomics. The posture detection also works offline with the MSCOCO dataset specifically on the key point task (see: https://cocodataset.org/keypoints-2020 (accessed on June 20, 2021). It is a public dataset that contains both gender details with 66,808 images, with people from different racial backgrounds and age groups (Piñero-Fuentes et al. 2021).

• Skeletal processing: TRT_Pose software was used to detect a human being’s real-time skeleton. The processing framework was based on CNN model which was also implemented in Pytorch (name: RestNet). RestNet uses an inertial connection between layers which are not contiguous to communicate information from earlier or later moments in time with respect to moment. Thus, information could be stored and reused for posture system classification. Finally, by using a deep learning model, the authors were able to detect the wearer’s joint position (defined by 18 points) (Piñero-Fuentes et al. 2021).
  
• Posture recognition: posture recognition by a posture detector uses wearable or sensor-based technology to monitor and assess an individual’s posture in real time. These devices, often in the form of sensors attached to clothing or worn on the body, track key indicators such as spinal alignment, body positioning and movement patterns. By collecting data from the joints and muscles, the posture detector can identify poor posture, such as slouching or misalignment, and provide immediate feedback through alerts or notifications. The aim of posture recognition, through the use of these devices, is to help users correct their posture, reduce the risk of musculoskeletal issues and promote better ergonomics, especially in environments such as workplaces or during rehabilitation. The system may also offer personalized guidance or reminders to maintain optimal posture throughout daily activities.

• Posture evaluation: posture evaluation using posture detectors has been the focus of several studies, which demonstrate the effectiveness of these devices in monitoring and improving body alignment. These studies typically use wearable sensors, pressure mats or vision-based systems to assess posture in real time, providing valuable insights into the body’s alignment during various activities. Researchers have found that posture detectors are particularly useful in identifying improper posture in both sedentary and dynamic movements, helping reduce musculoskeletal strain and discomfort. For instance, studies in workplace settings have used posture detectors to monitor employees’ sitting or standing positions, while clinical studies have applied them in rehabilitation for musculoskeletal disorders or after surgeries. The data collected from these devices allow for targeted interventions, such as posture correction techniques or ergonomic adjustments, significantly contributing to better posture awareness, prevention of posture-related injuries and overall health improvement (Piñero-Fuentes et al. 2021).