Dataset Information

A View Independent Classification Framework for Yoga Postures.

ABSTRACT: Yoga is a globally acclaimed and widely recommended practice for a healthy living. Maintaining correct posture while performing a Yogasana is of utmost importance. In this work, we employ transfer learning from human pose estimation models for extracting 136 key-points spread all over the body to train a random forest classifier which is used for estimation of the Yogasanas. The results are evaluated on an in-house collected extensive yoga video database of 51 subjects recorded from four different camera angles. We use a three step scheme for evaluating the generalizability of a Yoga classifier by testing it on (1) unseen frames, (2) unseen subjects, and (3) unseen camera angles. We argue that for most of the applications, validation accuracies on unseen subjects and unseen camera angles would be most important. We empirically analyze over three public datasets, the advantage of transfer learning and the possibilities of target leakage. We further demonstrate that the classification accuracies critically depend on the cross validation method employed and can often be misleading. To promote further research, we have made key-points dataset and code publicly available.

Supplementary information

The online version contains supplementary material available at 10.1007/s42979-022-01376-7.

SUBMITTER: Chasmai M

PROVIDER: S-EPMC9470077 | biostudies-literature | 2022

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A View Independent Classification Framework for Yoga Postures.

Chasmai Mustafa M Das Nirjhar N Bhardwaj Aman A Garg Rahul R

SN computer science 20220913 6

Yoga is a globally acclaimed and widely recommended practice for a healthy living. Maintaining correct posture while performing a Yogasana is of utmost importance. In this work, we employ transfer learning from human pose estimation models for extracting 136 key-points spread all over the body to train a random forest classifier which is used for estimation of the Yogasanas. The results are evaluated on an in-house collected extensive yoga video database of 51 subjects recorded from four differe ...[more]

PMID: 36120095

Similar Datasets

Project description:BACKGROUND:Most clinical investigations involving yoga lack adequate description of the specific yoga elements, including physical postures. Few studies have measured self-efficacy regarding the performance of yoga postures or assessed observed success in performing postures. METHODS:We developed and piloted several tools to evaluate self-efficacy and observed success in practicing yoga in the context of a randomized feasibility trial of an Iyengar-based yoga intervention for urinary incontinence in ambulatory women ?50?years. At the end of the 12-week yoga intervention involving twice weekly group yoga classes and once weekly home practice, participants rated their self-efficacy in performing each of the included 15 yoga postures on a 5-point Likert scale. During the 12th week, an expert yoga consultant observed participants and rated their competency in performing postures on a 5-point scale. Participants completed a questionnaire about self-efficacy in adhering to home yoga practice. We examined the distribution of and correlations between scores on the above measures. RESULTS:Among 27 participants (mean age 65?years), the range of means for self-efficacy ratings for individual postures was 3.6 to 4.5. The range of means for observed competency ratings for individual postures was 3.3 to 5.0. Mean self-efficacy rating for confidence in adhering to the assigned once-weekly home yoga practice was 2.8 (range 1 to 5). Posture self-efficacy was inversely correlated with participant age (p?=?0.01) and positively correlated with self-reported physical function (p?=?0.03) and mobility (p?=?0.01). No significant correlations were found between posture self-efficacy scale scores and expert-observed yoga competency ratings or practice adherence self-efficacy scores. CONCLUSIONS:These measures hold promise for advancing yoga research and practice by describing methods to: 1) measure self-efficacy in performing specific yoga postures; 2) use an expert observer to assess participants' competence in performing yoga postures; and 3) measure self-efficacy in adhering to home practice. These proposed measures can be used to describe specific components of yoga interventions, to assess whether study participants are able to learn to practice physical aspects of yoga and/or maintain this practice over time, as well as to investigate relationships between self-efficacy and competency in performing yoga postures to achieve specific health outcomes. TRIAL REGISTRATION:ClinicalTrials.gov, NCT02342678, January 21, 2015.

Project description:BackgroundView classification is a key step toward building a fully automated system for interpretation of echocardiograms. However, compared with adult echocardiograms, creating a view classification model for pediatric echocardiograms poses additional challenges, such as greater variation in anatomy, structure size, and views. The aim of this study was to develop a computer vision model to autonomously perform view classification on pediatric echocardiographic images.MethodsUsing a training set of 12,067 echocardiographic images from patients aged 0 to 19 years, a convolutional neural network model was trained to identify 27 preselected standard pediatric echocardiographic views which included anatomic sweeps, color Doppler, and Doppler tracings. A validation set of 6,197 images was used for parameter tuning and model selection. A test set of 9,684 images from 100 different patients was then used to evaluate model accuracy. The model was also evaluated on a per study basis using a second test set consisting of 524 echocardiograms from children with leukemia to identify six preselected views pertinent to cardiac dysfunction surveillance.ResultsThe model identified the 27 preselected views with 90.3% accuracy. Accuracy was similar across age groups (89.3% for 0-4 years, 90.8% for 4-9 years, 90.0% for 9-14 years, and 91.2% for 14-19 years; P = .12). Examining the view subtypes, accuracy was 78.3% for the cine one location, 90.5% for sweeps with color Doppler, 82.2% for sweeps without color Doppler, and 91.1% for Doppler tracings. Among the leukemia cohort, the model identified the six preselected views on a per study basis with a positive predictive value of 98.7% to 99.2% and sensitivity of 76.9% to 94.8%.ConclusionsA convolutional neural network model was constructed for view classification of pediatric echocardiograms that was accurate across the spectrum of ages and view types. This work lays the foundation for automated quantitative analysis and diagnostic support to promote efficient, accurate, and scalable analysis of pediatric echocardiograms.

Dataset Information

A View Independent Classification Framework for Yoga Postures.

Supplementary information

Publications

A View Independent Classification Framework for Yoga Postures.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets