Detection of sitting posture using hierarchical image composition and deep learning.

Audrius Kulikajevas,Rytis Maskeliunas,Robertas Damaševičius

doi:10.7717/peerj-cs.442

Audrius Kulikajevas, Rytis Maskeliunas + Show 1 more

Open Access

https://doi.org/10.7717/peerj-cs.442

Copy DOI

Abstract

Human posture detection allows the capture of the kinematic parameters of the human body, which is important for many applications, such as assisted living, healthcare, physical exercising and rehabilitation. This task can greatly benefit from recent development in deep learning and computer vision. In this paper, we propose a novel deep recurrent hierarchical network (DRHN) model based on MobileNetV2 that allows for greater flexibility by reducing or eliminating posture detection problems related to a limited visibility human torso in the frame, i.e., the occlusion problem. The DRHN network accepts the RGB-Depth frame sequences and produces a representation of semantically related posture states. We achieved 91.47% accuracy at 10 fps rate for sitting posture recognition.

Highlights

IntroductionAs our daily life increasingly depends on sitting work and the opportunities for physical exercising (in the context of COVID-19 pandemic associated restrictions and lockdowns are diminished), many people are facing various medical conditions directly related to such sedentary lifestyles
We propose a novel approach for human posture classification by using a supervised hierarchical neural network (Liu et al, 2019) that uses the RGB-Depth data as input
We have proposed an extension of the MobileNetV2 neural network, which allows the use of sequential video data as an input, allowing for the deep neural network to extract important temporal features from video frames, which would otherwise be lost when compared to a single-frame classification while still being capable of the single-frame prediction due to being biased towards the last frame

Summary

Introduction

As our daily life increasingly depends on sitting work and the opportunities for physical exercising (in the context of COVID-19 pandemic associated restrictions and lockdowns are diminished), many people are facing various medical conditions directly related to such sedentary lifestyles. One of the frequently mentioned problems is back pain, with bad sitting posture being one of the compounding factors to this problem (Grandjean & Hünting, 1977; Sharma & Majumdar, 2009). Inadequate postures adopted by office workers are one of the most significant risk factors of work-related musculoskeletal disorders. One study (Alberdi et al, 2018) has demonstrated that body posture is one of the best predictors of stress and mental workload levels. Another study linked postural instability and gait difficulty with a rapid rate of Parkinson’s disease progression

Methods

Results

Discussion

Conclusion