Real-time human sitting posture detection using mobile devices

Mobile Applications for Assessing Human Posture: A Systematic Literature Review

“…A total of 13 smartphone applications-RQ1 -for assessing human postural deviations were identified in this SLR and classified into applications based on the Use of Sensors (US) and Image-Based Solution (IBS) -RQ2. All mobile applications identified in the studies had an identification name, except that proposed by Estrada and Vea [65].…”

“…[66] Scoliosis [67] Rib Hump Angle/Scoliosis [68] Kyphotic Angles [69] Cobb Angle/Thoracic Scoliosis [70] Cobb Angle/Thoracic Scoliosis [71] Lumbar Lordosis [72] Rib Hump Angle/Scoliosis Evaluation [73] Cobb Angle of Thoracolumbar Scoliosis [74] Cobb Angle [75] Cobb Angle [76] Cobb Angle [77] Head, Ribcage, Shoulder, Hip and Knee Tilt [65] Inclination of the Human Spinal Points [78] SS and Lumbar Curve [79] PI, PT and LL [80] Sagittal Curvaures [81] Thoracic Spine Rotation [82] LL, PI, and PT [83] Head, Ribcage, Shoulder, Hip Tilt [84] Head, Ribcage, Shoulder, Hip Inclination [85] Scoliosis Angle [86] Head, Shoulder and Ribcage Tilt [87] Head, Ribcage, Shoulder, Hip and Knee Tilt [88] LL, PI, SS and PT [89] LL, PT, and SS [90] Rib Hump/Scoliosis Evaluation Note: LL = Lumbar Lordosis; PI = Pelvic Incidence; SS = Sacral Slope; PT = Pelvic Tilt.…”

“…Except for the adapters used in the studies of Chen et al [90] and Izatt et al [67], only in the study of Estrada and Vea [65] was an auxiliary device (a kind of adjustable vest) used as a means of fitting the smartphones to the participant's body. The aim is to ensure that mobile devices are correctly in contact with the three points (thoracic, thoracolumbar and lumbar) of the spine.…”

“…[77,83,84,86,87] Among the studies identified, only one reported use of machine learning algorithms in the proposed mobile application. Estrada and Vea [65] developed a solution to detect proper/improper sitting postures based on the angles of the spinal curvatures. To identify the best algorithm to perform this task, the authors conducted performance evaluations with the machine learning classifiers: k-Nearest Neighbors (KNN), Support-vector Machines (SVM), Multilayer Perceptron (MLP), a class of artificial neural network, and Decision Tree.…”

“…Not all studies included in this review made comparisons between the studied/proposed mobile applications and other postural assessment methods. Some of them [65,77,[83][84][85][86] evaluated the ability of mobile applications to measure postural deviations, but without comparisons with other tools or traditional methods used for postural assessment. Analyses, comparative or not with other methods, were based on a varied number of statistical analysis, as indicated in Table 11.…”