A computer vision-based mobile tool for assessing human posture: A validation study

Vision-based postural balance assessment of sit-to-stand transitions performed by younger and older adults

BACKGROUND

The use of inertial measurement units (IMUs) in assessing fall risk is often limited by subject discomfort and challenges in data interpretation. Additionally, there is a scarcity of research on attitude estimation features. To address these issues, we explored novel features and representation methods in the context of sit-to-stand transitions. This study recorded sit-to-stand transition test data from three groups: community-dwelling elderly, elderly in day care centers (DCC), and college students, captured using mobile phone cameras.

METHOD

We employed pose estimation technology to extract key point kinematic features from the video data and used 10-fold cross-validation to train a random forest classifier, mitigating the impact of individual differences. We trained classifiers with the top 5, 10, and 15 features, calculating the average area under the receiver operating characteristic curve (AUC) for each model to compare feature importance.

RESULTS

Our results indicated that elbow key point features, such as (KP08) mean Y, (KP08)RMS Y, (KP09) mean Y, and (KP09) RMS Y, are crucial for distinguishing between subject groups. Statistical tests further validated the significance of these features. The application of human pose estimation and key point signals shows promise for clinical postural balance screening. The identified features can be utilized to develop non-invasive tools for assessing postural instability risk, contributing to fall prevention efforts.

CONCLUSION

This study lays the groundwork for integrating additional measurement modalities into sit-to-stand transition analysis to enhance clinical strategies.

Validity Analysis of Monocular Human Pose Estimation Models Interfaced with a Mobile Application for Assessing Upper Limb Range of Motion

Human Pose Estimation (HPE) is a computer vision application that utilizes deep learning techniques to precisely locate Key Joint Points (KJPs), enabling the accurate description of a person's pose. HPE models can be extended to facilitate Range of Motion (ROM) assessment by leveraging patient photographs. This study aims to evaluate and compare the performance of HPE models for assessing upper limbs ROM. A physiotherapist evaluated the degrees of ROM in shoulders (flexion, extension, and abduction) and elbows (flexion and extension) for fifty-two participants using both Universal Goniometer (UG) and five HPE models. Participants were instructed to repeat each movement three times to obtain measurements with the UG, then positioned while photos were captured using the NLMeasurer mobile application. The paired t-test, bias, and error measures were employed to evaluate the difference and agreement between measurement methods. Results indicated that the MoveNet Thunder INT16 model exhibited superior performance. Root Mean Square Errors obtained through this model were <10° in 8 of 10 analyzed movements. HPE models demonstrated better performance in shoulder flexion and abduction movements while exhibiting unsatisfactory performance in elbow flexion. Challenges such as image perspective distortion, environmental lighting conditions, images in monocular view, and complications in the pose may influence the models' performance. Nevertheless, HPE models show promise in identifying KJPs and facilitating ROM measurements, potentially enhancing convenience and efficiency in assessments. However, their current accuracy for this application is unsatisfactory, highlighting the need for caution when considering automated upper limb ROM measurement with them. The implementation of these models in clinical practice does not diminish the crucial role of examiners in carefully inspecting images and making adjustments to ensure measurement reliability.

A comprehensive analysis of the machine learning pose estimation models used in human movement and posture analyses: A narrative review

The accurate measurement and analysis of human movement are essential in fields ranging from rehabilitation and neuroscience to sports science and ergonomics. Traditional methods, though precise, are often constrained by cost, accessibility, and controlled environments. The advent of machine learning (ML) pose estimation models (PEMs) offers an alternative solution, enabling detailed motion analysis using low-cost imaging systems in various settings. The aim of this review is to evaluate ML PEMs and their impact on human movement sciences, focusing on recent advancements in machine learning and computer vision for accurate, non-invasive motion analysis using low-cost imaging systems. A narrative review was conducted by searching electronic databases, including PubMed and Google Scholar, using key terms such as "machine learning," "pose estimation models," and "human movement sciences." Thematic analysis identified key advancements, applications, and challenges in ML PEMs across clinical, sports, and ergonomic contexts. The review highlights the development, capabilities, and applications of models such as OpenPose, PoseNet, AlphaPose, DeepLabCut, HRNet, MediaPipe Pose, BlazePose, EfficientPose, and MoveNet, emphasizing their potential for non-invasive, cost-effective assessments. In clinical settings, these models enable objective gait and posture analysis, aiding in diagnosing musculoskeletal disorders and tracking rehabilitation progress. In sports, ML PEMs enhance performance analysis and injury prevention by providing real-time feedback and detailed biomechanical data. In ergonomics, they offer proactive solutions for workplace injury prevention through real-time posture and movement analysis. While promising, the implementation of ML PEMs faces challenges in accuracy, data quality, and integration into existing practices. Establishing standardized protocols and frameworks is crucial for ensuring reliable, interdisciplinary applications. This review can be useful for coaches, healthcare professionals, and researchers in evaluating and implementing ML PEMs, ultimately advancing the field of human movement sciences.

Use of Artificial Intelligence in Cobb Angle Measurement for Scoliosis: Retrospective Reliability and Accuracy Study of a Mobile App

BACKGROUND

Scoliosis is a spinal deformity in which one or more spinal segments bend to the side or show vertebral rotation. Some artificial intelligence (AI) apps have already been developed for measuring the Cobb angle in patients with scoliosis. These apps still require doctors to perform certain measurements, which can lead to interobserver variability. The AI app (cobbAngle pro) in this study will eliminate the need for doctor measurements, achieving complete automation.

OBJECTIVE

We aimed to evaluate the reliability and accuracy of our new AI app that is based on deep learning to automatically measure the Cobb angle in patients with scoliosis.

METHODS

A retrospective analysis was conducted on the clinical data of children with scoliosis who were treated at the Pediatric Orthopedics Department of the Children's Hospital affiliated with Fudan University from July 2019 to July 2022. Three measurers used the Picture Archiving and Communication System (PACS) to measure the coronal main curve Cobb angle in 802 full-length anteroposterior and lateral spine X-rays of 601 children with scoliosis, and recorded the results of each measurement. After an interval of 2 weeks, the mobile AI app was used to remeasure the Cobb angle once. The Cobb angle measurements from the PACS were used as the reference standard, and the accuracy of the Cobb angle measurements by the app was analyzed through the Bland-Altman test. The intraclass correlation coefficient (ICC) was used to compare the repeatability within measurers and the consistency between measurers.

RESULTS

Among 601 children with scoliosis, 89 were male and 512 were female (age range: 10-17 years), and 802 full-length spinal X-rays were analyzed. Two functionalities of the app (photography and photo upload) were compared with the PACS for measuring the Cobb angle. The consistency was found to be excellent. The average absolute errors of the Cobb angle measured by the photography and upload methods were 2.00 and 2.08, respectively. Using a clinical allowance maximum error of 5°, the 95% limits of agreement (LoAs) for Cobb angle measurements by the photography and upload methods were -4.7° to 4.9° and -4.9° to 4.9°, respectively. For the photography and upload methods, the 95% LoAs for measuring Cobb angles were -4.3° to 4.6° and -4.4° to 4.7°, respectively, in mild scoliosis patients; -4.9° to 5.2° and -5.1° to 5.1°, respectively, in moderate scoliosis patients; and -5.2° to 5.0° and -6.0° to 4.8°, respectively, in severe scoliosis patients. The Cobb angle measured by the 3 observers twice before and after using the photography method had good repeatability (P<.001). The consistency between the observers was excellent (P<.001).

CONCLUSIONS

The new AI platform is accurate and repeatable in the automatic measurement of the Cobb angle of the main curvature in patients with scoliosis.

Biomechanical Posture Analysis in Healthy Adults with Machine Learning: Applicability and Reliability

Posture analysis is important in musculoskeletal disorder prevention but relies on subjective assessment. This study investigates the applicability and reliability of a machine learning (ML) pose estimation model for the human posture assessment, while also exploring the underlying structure of the data through principal component and cluster analyses. A cohort of 200 healthy individuals with a mean age of 24.4 ± 4.2 years was photographed from the frontal, dorsal, and lateral views. We used Student's t-test and Cohen's effect size (d) to identify gender-specific postural differences and used the Intraclass Correlation Coefficient (ICC) to assess the reliability of this method. Our findings demonstrate distinct sex differences in shoulder adduction angle (men: 16.1° ± 1.9°, women: 14.1° ± 1.5°, d = 1.14) and hip adduction angle (men: 9.9° ± 2.2°, women: 6.7° ± 1.5°, d = 1.67), with no significant differences in horizontal inclinations. ICC analysis, with the highest value of 0.95, confirms the reliability of the approach. Principal component and clustering analyses revealed potential new patterns in postural analysis such as significant differences in shoulder-hip distance, highlighting the potential of unsupervised ML for objective posture analysis, offering a promising non-invasive method for rapid, reliable screening in physical therapy, ergonomics, and sports.

Concurrent validity and reliability of a smartphone-based application for the head repositioning and cervical range of motion

Objectives

To evaluate the validity and reliability of a smartphone-based application against inertial sensors to measure head repositioning (by using joint position sense –JPS) and cervical range of motion (ROM).

Methods

JPS and cervical ROM were evaluated for neck flexion, extension and both-sides lateral flexion in thirty-one volunteers. Participants were simultaneously evaluated with inertial sensors and the smartphone application. A total of 248 angles were compared for concurrent validity. Inter-tester and intra-tester reliability were evaluated through scoring of images with the smartphone application by two testers, and re-scoring images by the same tester.

Results

Very high correlation was observed between both methods for ROM in all neck movements and JPS in left-side lateral flexion (r>0.9), and high for JPS in the rest of movements (r>0.8). Bland-Altman plots always demonstrated absolute agreement. Inter-and intra-tester reliability was perfect for JPS and ROM in all the neck movements (ICC>0.81).

Conclusions

This smartphone-based application is valid and reliable for evaluating head repositioning and cervical ROM compared with inertial sensors in healthy and young adults. Health professionals could use it in an easier and portable way in field conditions.

Front-end deep learning web apps development and deployment: a review

Machine learning and deep learning models are commonly developed using programming languages such as Python, C++, or R and deployed as web apps delivered from a back-end server or as mobile apps installed from an app store. However, recently front-end technologies and JavaScript libraries, such as TensorFlow.js, have been introduced to make machine learning more accessible to researchers and end-users. Using JavaScript, TensorFlow.js can define, train, and run new or existing, pre-trained machine learning models entirely in the browser from the client-side, which improves the user experience through interaction while preserving privacy. Deep learning models deployed on front-end browsers must be small, have fast inference, and ideally be interactive in real-time. Therefore, the emphasis on development and deployment is different. This paper aims to review the development and deployment of these deep-learning web apps to raise awareness of the recent advancements and encourage more researchers to take advantage of this technology for their own work. First, the rationale behind the deployment stack (front-end, JavaScript, and TensorFlow.js) is discussed. Then, the development approach for obtaining deep learning models that are optimized and suitable for front-end deployment is then described. The article also provides current web applications divided into seven categories to show deep learning potential on the front end. These include web apps for deep learning playground, pose detection and gesture tracking, music and art creation, expression detection and facial recognition, video segmentation, image and signal analysis, healthcare diagnosis, recognition, and identification.

Postural Evaluation in Young Healthy Adults through a Digital and Reproducible Method

Different tools for the assessment of posture exist, from the simplest and cheap plumb line to complex, expensive, 3D-marker-based systems. The aim of this study is to present digital postural normative data of young adults collected through a mobile app to expand the possibilities of digital postural evaluation. A sample of 100 healthy volunteers, 50 males and 50 females, was analyzed with the mobile app Apecs-AI Posture Evaluation and Correction System® (Apecs). The Student’s t-test evaluated differences between gender to highlight if the digital posture evaluation may differ between groups. A significant difference was present in the anterior coronal plane for axillary alignment (p = 0.04), trunk inclination (p = 0.03), and knee alignment (p = 0.01). Head inclination (p = 0.04), tibia shift (p = 0.01), and foot angle (p < 0.001) presented significant differences in the sagittal plane, while there were no significant differences in the posterior coronal plane. The intraclass correlation coefficient (ICC) was considered to evaluate reproducibility. Thirteen parameters out of twenty-two provided an ICC > 0.90, three provided an ICC > 0.60, and six variables did not meet the cut-off criteria. The results highlight that digital posture analysis of healthy individuals may present slight differences related to gender. Additionally, the mobile app showed good reproducibility according to ICC. Digital postural assessment with Apecs could represent a quick method for preventing screening in the general population. Therefore, clinicians should consider this app’s worth as an auxiliary posture evaluation tool.