Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time

Posture Correction Application

Kushal Ganesh

and Amar Ramudhin

PhD candidate, Harrisburg University, 326 Market St, Harrisburg, PA 17101, Pennsylvania, U.S.A.

Professor and Program Lead of Information Systems Engineering and Management, Harrisburg University, 326 Market St,

Harrisburg, PA 17101, Pennsylvania, U.S.A.

Keywords: Yoga, Artificial Intelligence, Real-Time Alter System, Refined Learning, Neural Networks.

Abstract: The extensive propagation of yoga, once reserved for isolated spiritual communities, has brought several

advantages as more and more people have gained access to a holistic well-being system. Nonetheless, while

many individuals have benefited from the influence of yoga, lax application of its various regimens has

resulted in a hike in yoga-related injuries. The following paper explores current AI tools for yoga and then

elucidates the physical, psychological, and long-term adverse effects of improper yoga practice, with

supporting data and statistics. Furthermore, it suggests an AI-equipped application that alters the potential of

injuries arising from incorrect body movements during yoga. Next, particular technical features of the app are

enumerated, demonstrating how machine learning aids in data analytics to prevent misapplication and take

care of user privacy issues. Finally, the challenges of the proposed solution's adoption, usability, and ethical

usage are addressed, and suggestions are offered to circumvent such hurdles.

1 INTRODUCTION

Yoga originated in India over 2,000 years ago, and it

is now a primary global industry and a kaleidoscopic

cultural phenomenon enjoyed by hundreds of

millions worldwide. In 2020, the Yoga Alliance, the

US's most prominent non-profit association

representing the yoga community, found that almost

36 million people in the US practice yoga. Annual

growth since 2016 has been close to 10 percent.

However, as global yoga continues to expand, the

number of yoga injuries is also rising. One study

found that the total number of injuries treated at

emergency departments in the US between 2014 and

2017, with a primary diagnosis related to a yoga

practice, increased by 45 percent. Three of the most

common yoga-related injuries are sprains, strains, and

fractures.

The physiological, mental, and chronic harmful

consequences of incorrect yoga practice are discussed

in detail. Then, an AI-based yoga application is

proposed will be provided real time alerts to correct

posture utilising data analytics and machine learning.

2 YOGA AND AI

Yoga and AI are increasingly merging, combining

yoga's mindfulness and physical discipline with the

technological advancements of artificial intelligence.

This convergence allows practitioners to enhance

their routines, personalize experiences, and make

yoga more accessible.

2.1 AI-Powered Yoga Apps

With apps that provide real-time feedback during a

home-based, step-by-step yoga practice routine, AI is

impacting how yoga is performed as a discipline. Some

popular AI-powered apps include Skill Yoga

Movement Coach, which uses AI and computer vision

to track the practitioner's body alignment in real-time.

It provides feedback on yoga poses, helping

practitioners improve posture and form without

needing a physical instructor. Alo Moves Powered by

AI, the app recommends yoga classes based on

progress, fitness goals, and other user preferences.

QuickPose provides Real-time feedback on pose

accuracy and promotes user confidence and safety

while practicing. QuickPose is an AI-powered pose

Ganesh, K. and Ramudhin, A.

Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time Posture Correction Application.

DOI: 10.5220/0013063300003828

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 12th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2024), pages 241-250

ISBN: 978-989-758-719-1; ISSN: 2184-3201

241

detection software that gives users real-time feedback

on pose accuracy and helps improve a user's pose.

However, these existing tool are subscription based

and would put it out of reach of a huge population

segment.

2.2 AI in Wearable Technology

AI has also penetrated wearable technology, offering

an additional analytical layer for yoga practice.

Fitness trackers and other body-worn technology can

read the practitioner's physiological data (heart rate,

breathing patterns) and, by using AI, suggest tweaks

or the best times to practice. For instance, exercise

trackers such as Fitbit and Apple Watch, which make

use of sensors like accelerometers, gyroscopes, heart

rate monitors, and photo-optic sensors, among others,

include AI-powered yoga tracking features to help

assess poses in real-time, provide users with periodic

feedback on their progress over time, and provide

personalized coaching recommendations. The

efficacy of these apps would depend on ethical and

responsible data usage in line with the existing data

governing principles.

2.3 Virtual Yoga Instructors

AI is used to create virtual instructors that guide users

through yoga sessions, offering convenience and real-

time personalization. Yoga is no longer limited to set

time slots since users can connect with virtual yoga

instructors and spontaneously begin their practice at

any time of the day or night. Incorporating the

flexibility of technology, the AI-powered app Yogaia

allows one to connect with a live virtual class that

adapts to your improvement.

2.4 Mental Health and Meditation

Support

AI enhances mental wellness in yoga by offering

personalized meditation and mindfulness practices.

Apps such as Calm and Headspace give users tailored

meditation and mindfulness sessions using AI that

learns a user's mood and responds and adjusts a session

based on biometric signals and a user's rating of their

stress levels. These apps use AI to provide a meditation

or breathing exercise based on data and user inputs,

making the mental benefits of yoga more attainable.

2.5 AI in Yoga Studios

AI is being used not only to support individual

practice but also to enhance the in-studio experience.

For example, with AI, a studio may be able to shape

class settings—such as the lighting or temperature—

based on the preferences of whoever is walking

through the door. Alternatively, AI can help teachers

better support learners by identifying their needs

based on past injuries or class goals.

2.6 Yoga Therapy and Rehabilitation

Tools are being developed for yoga therapy and

rehabilitation that assist people in recovering from

injuries, tracking their recovery, and offer suggested

yoga poses to promote rehabilitation while informing

the practitioners of what to avoid to avoid further

injury.

However, AI cannot replace the human

connection with other practitioners and teachers that

yoga brings. AI tools cannot replace the more in-

depth and spiritual experiences that have attracted

people to yoga. Like in other domains, many

practitioners and experts have jumped onto the

bandwagon of the innovations in AI tools as methods

to assist bigger audiences in traditional yoga teaching

rather than replacing it. In conclusion, the great

benefits that AI is bringing to yoga are opening up

new horizons in terms of accessibility,

personalization, and ease of practice.

3 THE PHYSICAL AND MENTAL

RISKS OF IMPROPER YOGA

PRACTICE

3.1 Physical Injuries: Data and

Statistics

While yoga has established its place in the world of

exercise for its potential to improve flexibility,

muscle tone, and mental well-being, injury from poor

practice is sadly a regular occurrence. Research

published in the Journal of Bodywork and Movement

Therapies in 2018 reveals that, of the yoga students

they surveyed, 20 percent had, at some time, suffered

at least one injury severe enough to disrupt their

regular practice. Here are the common injuries:

3.1.1 Muscle Strains and Sprains

Overextension in forward bends, Hamstring injuries,

and splits are some joint muscle strains and sprains

caused by improper yoga practice. For example, a

study published in the American Journal of Science

icSPORTS 2024 - 12th International Conference on Sport Sciences Research and Technology Support

242

and Medicine in Sports concluded that 33 percent of

injuries sustained in yoga were hamstring strains.

3.1.2 In Joints

Weight-bearing poses such as downward dog and

handstands put your wrists, shoulders, and knees at

risk of injury. A 2017 study found that 21 percent of

yoga injuries affect the upper extremities more than

any other body part, primarily due to compression and

faulty alignment.

3.1.3 Spine Problems

Spinal and back injuries such as herniated discs and

sciatica can result from the incorrect execution of

spinal twists or back-bending poses. A study in the

International Journal of Yoga found that 10 percent of

practitioners reported chronic lower back pain from

yoga practice, pointing to improper technique.

3.1.4 Balance-Related Injuries

Balance poses (like tree poses or headstands)

commonly have poor alignment and often cause a fall.

The British Journal of Sports Medicine suggests that

balance-related injuries account for 14 percent of all

yoga injuries and are often associated with fractures

or concussions.

3.2 Mental and Emotional Strain

When practiced appropriately, yoga can make the

practitioner mentally and physically light and supple,

but it can have numerous undesirable emotional and

mental impacts if practiced wrongly.

3.2.1 Stress and Anxiety

When poses are held and advanced poses are

performed, both of which can involve exerting

physical strength and strain, stress and its

companions, anxiety – mainly focussed upon

performance – can result. A 2019 paper in Brain

plasticityreported that 15 percent of them found that

the performance pressures led to a rise in their

feelings of anxiety.

3.2.2 Mental Suffering

Misalignment and discomfort could cause a type of

mental fatigue, reducing the intended meditative

benefits. A study in the Evidence-based

Complementary and Alternative Medicine showed

that 25 percent of practitioners reported having lower

Mindfulness, becoming less focused, and getting less

relaxation purported to come from meditation

practice, attributable to poor alignment due to

inadequate instruction.

3.3 Long-Term Bodily Harm

Not only can the aftershock of a poorly practiced or

arbitrary pose result in acute injury, but it can also

take years to manifest as a chronic disease.

3.3.1 Misalignment

Pelvic misalignment, in particular, can lead to pain in

the lower back and neck. An article in The Journal of

Yoga and Physical Therapy reported that one-third of

yoga practitioners who had been practicing for an

extended period developed chronic pain as a result of

misalignment.

3.3.2 Postural Imbalances

Chronic incorrect practice may also cause postural

issues, such as the development of muscular

imbalances. One study from 202 found that 22 percent

of practitioners developed such musculoskeletal

imbalances from chronic incorrect postures.

3.3.3 Neurological Damage

Poses involving neck or head misalignment could

cause neurological damage. Rare, severe injuries to

the neck area, to specific nerves, and even stroke have

been reported. For instance, The Public Library of

Science published cases of nerve compression due to

inappropriate headstands.

3.4 The Role of Technology in

Preventing Improper Yoga Practice

As improper yoga practice is becoming common, we

need technological solutions that can help reduce

injury by providing real-time feedback on the asanas

and giving immediate corrective feedback. This paper

proposes an Artificial Intelligence (AI) based yoga app

that can track a person's posture and give feedback

when required. The difficulty in practicing various

asanas increases due to the excessive burden of

professional work, unhealthy lifestyles, and eating

habits, causing many youngsters to develop weak

physical conditions. These factors are leading to cases

of improper yoga and unintended injuries. Yoga is a

beneficial practice in daily life, resulting in holistic

wellness. We propose an AI-based mobile yoga app to

address the abovementioned problems. The app can

Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time Posture Correction Application

243

alert and predict people's movements while performing

various asanas. The algorithm implemented in the app

enables it to predict the outcome of practice and give

immediate feedback when needed. During the practice,

the app's AI will constantly keep track of the posture.

While it will allow users to proceed when the expected

posture is achieved, it will intervene if the user's body

posture is incorrect. The proposed app not only alters

the incorrect yoga poses but also provides virtual

assistance if needed, as well as initiating emergency

protocol if a severe potential injury persists. Hence, this

will reduce the issues and injuries from improper yoga

practice.

4 KEY FEATURES OF THE

AI-POWERED YOGA

APPLICATION

4.1 Real-Time Pose Estimation

Behind this, computer vision and pose estimation

algorithms are used to track with the the proposed

solution enhances mobile device's camera and

compare it with ones from a database. At the heart of

the app is its pose estimation system, which relies on

computer vision algorithms such as OpenPose or

MoveNet to detect and track landmarks (such as

shoulders, hips, knees, and ankles) on the human

body in real time.

The app relies upon a convolutional neural

network (CNN) trained on a sizeable open-source

dataset of varied yoga poses. This allows it to

accurately identify many of humans' essential body

landmarks, such as joints, angles, and muscle gears.

4.2 Personalized Feedback and

Correction

After the practitioner assumes a pose, the app

immediately explains which areas they must correct

until their body is aligned correctly. For instance, if

the practitioner is practicing the cobra pose and their

spine is not straight, the app will alert them and guide

them to straighten it better.

The app uses data analytics processing and

extraction to interpret user data and transform the raw

posture at a specific time into user-actionable

insights. With machine learning models specially

built for supervised learning, the posture self-

assessment app detects patterns in the wrong posture

and predicts posture risks.

It tailors its feedback to the practitioner's unique

body and the specific history of practice found in its

database.

4.3 Data-Driven Insights and Progress

Tracking

Once installed, the application takes a lifetime

inventory of the user's posture as they practice,

crowdsourcing the data and analyzing it for repetitive

tendencies and errant postural quirks over the long

term. The data is combined with knowledge about the

standard movements that lead to repetitive strain and

cramping to generate personalized practice plans and

workout routines tailored to the user's particular areas

of weakness. The app uses reinforcement learning

techniques, in part, to calibrate the feedback and

practice routines. The more it learns about the user, the

better the app becomes at helping the user improve.

A dashboard will allow individuals to track their

progress via visual reports, such as improvements in

skeletal alignment, range of motion, and other

assessments of body posture quality.

4.4 AI-Driven Personalization

The app continuously becomes more adept at making

posture corrections using machine learning. The app

learns about the user's practice style – the more the user

works with the app, the more refined the AI can

become in making corrections and improving overall

performance.

Additionally, users get their yoga instruction

tailored to their objectives by selecting their target

(flexibility, strength, balance, and so on), with each

workout catered accordingly.

5 TECHNICAL FLOW

5.1 Overall Flow

5.1.1 Input Video Data

This is the stage that will handle the video data, which

processes the video feed from the user's mobile

camera. The characteristics of the input such as

encoding format (e.g., MP4, AVI), resolution

(e.g.,1080p, 4K), and frame rate (e.g., 30fps, 60fps),

are analysed and normalised if needed.

5.1.2 Extract Metadata

In the next step, the metadata is extracted, such as the

file name, video and audio duration, video and audio

icSPORTS 2024 - 12th International Conference on Sport Sciences Research and Technology Support

244

codecs (e.g., H.264, AAC), aspect ratio, frame rate,

and bit rate, among other features. This metadata is

important because it informs various aspects of the

preprocessing and AI modeling so that the video is

transformed in ways appropriate to its specific form.

5.1.3 Preprocessing Stage

The preprocessing step, as the first stage of the data

pipeline, is fundamental for the AI model’s input as it

consists of a few sub-processes:

• Frame Extraction: Each frame in the video is

separated from the others.

• Resolution Adjustment: The decomposed video

stream comprises extracted frames that are then

rescaled to the target resolution and aspect ratio.

As a result, the output complies with both its

visual quality and the resolution requirements

set by its use application.

• Noise Reduction: This subprocess involves

applying noise filters to the frames to eliminate

unwanted noises visually, improving the quality

of the video.

• Colour Correction: Brightness, contrast,

saturation, and hue, among other frame

characteristics, are adjusted so that shots are

consistent.

5.1.4 AI Model Input Data Preparation

After preprocessing, the video data is prepared for AI

model input through several techniques:

• Feature Extraction: Extracting salient features like

edge, corner, texture, interest points, and motion

vectors from the video frames. All these features

form the essential ingredients of the AI model to

understand what the video contains.

• Object Detection: The algorithm recognizes

objects within the video and assigns them to a

set of pre-defined classes to help extract

meaning from the video.

• Scene Segmentation: The video is divided into

multiple regions and scenes, which can help the

AI model improve its predictions on each

segment because each region represents a

different activity of the yoga practitioner.

5.1.5 AI Model Processing

At this point, the (preprocessed) video stream data are

fed into the AI model, which is used to train model

weights, neural network architecture parameters, and

inference parameters to perform AI-driven

enhancements and analyses such as object recognition,

action detection, and scene understanding.

5.1.6 Post-Processing Stage

Following the AI model's processing, the video

undergoes post-processing:

• Overlay Annotations: The video is shown with

detected objects, labels, and tracking data. This

step transforms the event video by adding

annotations from the AI models, allowing the

views to customise the view.

• Rendering the final frames: This is the final step

where everything gets knitted together for the

user display.

5.1.7 Output Generation

The last stage involves generating the final outputs:

• Deliver final Video: The processed video is

displayed in real-time.

• Logs/reports: The system logs and reports

processing steps, errors, and performance

metrics in detail. This is critical for

understanding performance, debugging errors,

and maintaining the system.

5.2 User Alert System

5.2.1 Capture Video Data

When the session starts, the mobile starts to capture

the user practising yoga. This captured data enables

constant estimation of the posture during the entire

session.

5.2.2 Analyze Posture

This is where the AI takes the incoming live video

feed and runs it through computer vision and

algorithms to determine critical points on the user's

body—for instance, the shoulders, the hips, the

ankles, etc. Once the various parts of the user’s body

are successfully identified, the AI system compares

the user’s ‘live’ posture with pre-trained ideal

postures in the AI’s database. The basis of these ideal

postures is standard yogic positions from open-

sourced data carefully calibrated by human experts

such as yoga teachers.

5.2.3 Posture Assessment

In the first stage, the AI analyses the posture. If the

posture is correct, the system monitors it continuously

Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time Posture Correction Application

245

Figure 1: Overall technical flow.

without taking any action. However, if the AI

understands that the posture is incorrect, the AI will

alert the user, and the user is expected to correct their

posture. The alert is sent via visual. However, the

auditory alert settings will be dependent on the user's

needs and preferences.

5.2.4 Reassessment

After the alert, the system checks whether the posture

has been corrected. If the user has corrected the

posture, the system turns back to the normal

monitoring state. In contrast, if the posture is still not

corrected, the risk level counter, which counts the

number of times that the user has not corrected his/her

posture, is incremented. This counter is used for risk

assessment over a prolonged period of bad posture

during the session.

5.2.5 Risk Level Check

There is an active monitoring of the user’s posture

with the consequent counter of the risk level. If the

risk level is not high, the system continues monitoring

with periodic alerts. If the risk level exceeds a limit,

the system suggests that the user take a break to

prevent damage due to prolonged maintenance of the

incorrect posture.

5.2.6 Continuous Risk Check

After a break, the system checks whether the user

resumes an unsafe posture. If the user maintains the

correct position, the system continues the monitoring.

However, if the user reverts to continuing with the

unsafe posture, the risk counter increases again,

which makes it more probable that the level of

intervention escalates.

5.2.7 Severe Risk Assessment

If the risk counter rises beyond the threshold, the

system takes more elevated measures by connecting

the user to a virtual yoga instructor. Ideally, the

instructor should offer tailored recommendations to

help the user improve posture and perform the

icSPORTS 2024 - 12th International Conference on Sport Sciences Research and Technology Support

246

Figure 2: Alert system technical flow.

physical activity/exercise safely and more accurately.

Normally, it is difficult for a person receiving a

program guided by textual or video instructions to

adjust their behavior because they cannot quickly or

accurately receive feedback. The virtual instructor

can promptly respond to the user’s activity and

outcomes to facilitate behavior change.

5.2.8 Emergency Situation Check

If extreme issues are detected, such as those that could

harm the user, the system deploys an emergency dial

sequence. This action could be a call to 911 or some

other predetermined emergency procedure.

5.2.9 End

The session ends only when the user stops or is halted

due to a high-risk signal by the system at the end of

the exercise routine. The app saves the collected data

for future analysis and helps users compare and

optimize their posture over time.

6 CHALLENGES

6.1 Adoption and Reach

Even though the majority of people around the world

now have access to a smartphone, reliable access to

suitable hardware and stable internet is limited in

some areas. For instance, Indian women are 15

percent less likely to own a mobile phone and 33

percent less likely to use mobile internet services than

men. Furthermore, only 31 percent of the rural

population uses the Internet compared to 67 percent

of their urban counterparts. The variance in culture,

language, and dialects must all be factored in, which

would substantially add to execution cost and time.

6.2 Technical Limitations

Real-time video processing apps could face several

limitations due to computational complexity, data

throughput, and system resource constraints. High

computational demands for tasks such as object

detection, motion tracking, or image enhancement can

strain mobile device processors, often leading to

performance lags and overheating issues.

Predominantly, apps rely on low-latency data

processing, but maintaining frame rate and resolution

in real-time scenarios can be challenging, especially

with limited processing power and memory on mobile

devices.

Bandwidth limitations also pose a significant

challenge. Real-time video applications require

substantial data throughput, particularly for high-

definition streams, which can cause delays or reduced

video quality when network conditions are suboptimal.

Additionally, battery consumption is another concern,

as continuous real-time processing rapidly depletes

power in mobile environments.

6.3 User Experience and Engagement

To help users stay consistent with their practice, the

app employs tactics like gamification (earning

badges, etc.) and awarding rewards for things such as

consistently completing practice sessions every few

days for a specified number of times.

The app also has built-in social functions that

enable users to share their journey with friends or join

community challenges, which help create a sense of

community and motivate the user.

Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time Posture Correction Application

247

6.4 Ethical Considerations and Data

Privacy

The collection and administration of biometric data

raise ethical issues around consent and privacy. The

app should address these through full disclosure and

user control over data, following the GDPR.

The data is pseudonymized and encrypted; only

authorized personnel can access it. It should be

entirely under user control and disabled or deleted

whenever the user wants.

7 CONCLUSION

As yoga becomes more and more pervasive, not just as

a practice of wellness but also as a competitive sport,

the risks associated with the performance of improper

yoga increase as well. Our proposed AI yoga

application is a technological solution to these risks,

offering accurate analysis of the data of different users

for ideal yoga practice. The application relies on the

latest AI and machine learning applications, such as

deep learning, production, and pattern recognition

algorithms. Using the latest innovations in deep

learning, we have proposed an innovative solution to

help users avoid injuries and make their practice safer

and more effective. Hence, the proposed solution likely

enhances the benefits of AI systems for society and

ensures continued cultural respect for yoga. For such

technology to be accessible and usable globally,

challenges ranging from access, usability, and user-

interface issues to data privacy and cultural sensitivity

must be holistically addressed for the effective

implementation and usage of the proposed app.

REFERENCES

Wei, M. (2016). New survey reveals the rapid rise of yoga

— and why some people still haven’t tried it [Online].

Harvard Health Blog. Available from:

https://www.health.harvard.edu/blog/new-survey-

reveals-the-rapid-rise-of-yoga-and-why-some-people-

still-havent-tried-it-201603079179

Hessari, Mojedeh & TR, Prof. (2023). India's Pursuit Soft

Diplomacy of Intellectual Property Protection for Yoga,

Herbal Medicines, and Traditional Knowledge at the

WTO. 6. 3733 - 3748.

2016 Yoga in America Study Conducted by Yoga Journal

and Yoga Alliance Reveals Growth and Benefits of the

Practice | Yoga Alliance. (n.d.). Www.yogaalliance.org.

https://www.yogaalliance.org/Get_Involved/Media_In

quiries/2016_Yoga_in_America_Study_Conducted_by

_Yoga_Journal_and_Yoga_Alliance_Reveals_Growth

_and_Benefits_of_the_Practice

Sekendiz B. (2020). An epidemiological analysis of yoga-

related injury presentations to emergency departments

in Australia. The Physician and sportsmedicine, 48(3),

349–353. https://doi.org/10.1080/00913847.2020.

1717395

Skill Yoga GmbH. (2024). Skill Yoga: Yoga Workouts &

Personalized Training Plans. Skill Yoga. https://skill-

yoga.com/

New: (2021, July). Alo Moves. Alo Moves.

https://blog.alomoves.com/news/new-live-classes-on-

alo-moves

About Us - QuickPose.ai. (2024, May 13). QuickPose.ai.

https://quickpose.ai/about/

Shei, R. J., Holder, I. G., Oumsang, A. S., Paris, B. A., &

Paris, H. L. (2022). Wearable activity trackers-

advanced technology or advanced

marketing?. European journal of applied

physiology, 122(9), 1975–1990.

https://doi.org/10.1007/s00421-022-04951-1

Yogaia. (2021). Home Workouts Online for Every Body &

Mind | Yogaia.com. Yogaia.com. https://yogaia.com/

Calm. (2023). Experience Calm. Calm.

https://www.calm.com/

Headspace. (2023). Meditation and Mindfulness Made

Simple - Headspace. Headspace.

https://www.headspace.com/

Xplor Mariana Tek. (2023, September 19). Personalize

Your Yoga Studio with AI: A Comprehensive Guide &

10 Tools. Xplor Mariana Tek.

https://www.marianatek.com/blog/personalize-your-

yoga-studio-with-ai-a-comprehensive-guide-10-tools/

Embracing the Future: How AI is Revolutionizing Yoga

Practice - Doron Yoga. (2024, February 27).

https://doronyoga.com/embracing-the-future-how-ai-

is-revolutionizing-yoga-practice/

Campo M, Shiyko M, Kean M, Roberts L, Pappas E.

Musculoskeletal pain associated with recreational yoga

participation: A prospective cohort study with 1-year

follow-up. Journal of Bodywork and Movement

Therapies, Volume 22, Issue 2, 418 - 423.DOI:

10.1016/j.jbmt.2017.05.022

Carmer H, Ostermann T, Dobos G. Injuries and other

adverse events associated with yoga practice: A

systematic review of epidemiological studies. Journal

of Science and Medicine in Sport, Volume 21, Issue 2,

147 - 154. https://doi.org/10.1016/j.jsams.2017.08.026

Yoga more risky for causing musculoskeletal pain than you

might think: Injury rate up to 10 times higher than

previously reported. (n.d.). ScienceDaily.

https://www.sciencedaily.com/releases/2017/06/17062

7105433.htm

Colgrove, Yvonne M; Gravino-Dunn, Nicole S; Dinyer,

Sarah C; Sis, Emily A; Heier, Alexa C; Sharma, Neena

K. Physical and Physiological Effects of Yoga for an

Underserved Population with Chronic Low Back Pain.

International Journal of Yoga 12(3):p 252-264, Sep–

Dec 2019. | DOI: 10.4103/ijoy.IJOY_78_18

icSPORTS 2024 - 12th International Conference on Sport Sciences Research and Technology Support

248

CrevelÃ¡rio de Melo R, Victoria Ribeiro AÃ‚, Luquine Jr

CD, et alEffectiveness and safety of yoga to treat

chronic and acute pain: a rapid review of systematic

reviews. British medical journal Open

2021;11:e048536. doi: 10.1136/bmjopen-2020-048536

Gothe, N. P., Khan, I., Hayes, J., Erlenbach, E., &

Damoiseaux, J. S. (2019). Yoga Effects on Brain

Health: A Systematic Review of the Current Literature.

Brain plasticity (Amsterdam, Netherlands), 5(1), 105–

122. https://doi.org/10.3233/BPL-190084

Büssing, A., Michalsen, A., Khalsa, S. B., Telles, S., &

Sherman, K. J. (2012). Effects of yoga on mental and

physical health: a short summary of reviews. Evidence-

based complementary and alternative medicine: eCAM,

2012, 165410. https://doi.org/10.1155/2012/165410

Sawyer, Amy & Martinez, Sarah & Warren, Gordon.

(2012). Impact of Yoga on Low Back Pain and

Function: A Systematic Review and Meta-Analysis.

Journal of Yoga and Physical Therapy. 2. 120.

10.4172/2157-7595.1000120.

Susilowati IH, Kurniawidjaja LM, Nugraha S, Nasri SM,

Pujiriani I, Hasiholan BP. The prevalence of bad posture

and musculoskeletal symptoms originating from the use

of gadgets as an impact of the work from home program

of the university community. Heliyon.

2022;8(10):e11059.

doi:10.1016/j.heliyon.2022.e11059

Salsali M, Sheikhhoseini R, Sayyadi P, Hides JA, Dadfar

M, Piri H. Association between physical activity and

body posture: a systematic review and meta-analysis.

BMC Public Health. 2023;23(1):1670. Published 2023

Aug 30. doi:10.1186/s12889-023-16617-4

Koren, Leon & Stipancic, Tomislav & Ricko, Andrija &

Orsag, Luka. (2022). Person Localization Model Based

on a Fusion of Acoustic and Visual Inputs. Electronics.

11. 440. 10.3390/electronics11030440.

Huang, Xiaoyang & Lin, Zhi & Huang, Shaohui & Wang,

Fu Lee & Chan, Moon-Tong & Wang, Liansheng.

(2022). Contrastive learning–guided multi-meta

attention network for breast ultrasound video diagnosis.

Frontiers in Oncology. 12. 10.3389/fonc.2022.952457.

Gabriele Guarnieri, Marco Fontani, Francesco Guzzi,

Sergio Carrato, Martino Jerian.Perspective registration

and multi-frame super-resolution of license plates in

surveillance videos. Forensic Science International:

Digital Investigation, Volume 36,2021,301087, ISSN

2666-2817. https://doi.org/10.1016/j.fsidi.2020.

301087.

Mousa Al-Akhras, Zainab Darwish, Samer Atawneh,

Mohamed Habib. Improving Association Rules

Accuracy in Noisy Domains Using Instance Reduction

Techniques, Computers, Materials and Continua.

Volume 72, Issue 2, 2022, Pages 3719-3749, ISSN

1546-2218.

https://doi.org/10.32604/cmc.2022.025196.

Siqi Ye, Shao-Ping Lu, Adrian Munteanu. Color correction

for large-baseline multiview video. Signal Processing:

Image Communication. Volume 53, 2017. Pages 40-50,

ISSN 0923-5965. https://doi.org/10.1016/j.image.

2017.01.004.

Anima Pramanik, Sobhan Sarkar, Sankar K. Pal. Video

surveillance-based fall detection system using object-

level feature thresholding and Z−numbers, Knowledge-

Based Systems Volume 280, 2023, 110992, ISSN 0950-

7051. https://doi.org/10.1016/j.knosys.2023.110992.

Shreya Talati, Darshan Vekaria, Aparna Kumari, Sudeep

Tanwar. An AI-driven object segmentation and speed

control scheme for autonomous moving platforms,

Computer Networks Volume 186, 2021, 107783, ISSN

1389-1286. https://doi.org/10.1016/j.comnet.2020.

107783.

Calvin Perumalla, LaDonna Kearse, Michael Peven,

Shlomi Laufer, Cassidi Goll, Brett Wise, Su Yang, Carla

Pugh, AI-Based Video Segmentation: Procedural Steps

or Basic Maneuvers? Journal of Surgical Research,

Volume 283, 2023, Pages 500-506, ISSN 0022-4804.

https://doi.org/10.1016/j.jss.2022.10.069."

Nidhi Tewathia, Anant Kamath, P. Vigneswara Ilavarasan,

Social inequalities, fundamental inequities, and

recurring of the digital divide: Insights from India,

Technology in Society, Volume 61, 2020, 101251, ISSN

0160-791X. https://doi.org/10.1016/j.techsoc.2020.

101251.

Rahmillah FI, Tariq A, King M, Oviedo-Trespalacios O.

Evaluating the Effectiveness of Apps Designed to

Reduce Mobile Phone Use and Prevent Maladaptive

Mobile Phone Use: Multimethod Study. J Med Internet

Res. 2023;25:e42541. Published 2023 Aug 29.

doi:10.2196/42541

Qingyue Tan, Gerui Lv, Xing Fang, Jiaxing Zhang, Zejun

Yang, Yuan Jiang, and Qinghua Wu. 2024. Accurate

Bandwidth Prediction for Real-Time Media Streaming

with Offline Reinforcement Learning. In Proceedings

of the 15th ACM Multimedia Systems Conference

(MMSys '24). Association for Computing Machinery,

New York, NY, USA, 381–387.

https://doi.org/10.1145/3625468. 3652183

W. Zhang et al., "A Streaming Cloud Platform for Real-

Time Video Processing on Embedded Devices," in

IEEE Transactions on Cloud Computing, vol. 9, no. 3,

pp. 868-880, 1 July-Sept. 2021, doi:

10.1109/TCC.2019.2894621.

Isabelle Kniestedt, Iulia Lefter, Stephan Lukosch, Frances

M. Brazier. Re-framing engagement for applied games:

A conceptual framework, Entertainment Computing,

Volume 41, 2022, 100475, ISSN 1875-9521.

https://doi.org/10.1016/j.entcom.2021.100475."

Kai-Yu Wang, Abdul R. Ashraf, Narongsak Tek

Thongpapanl, Oanh Nguyen. Influence of social

augmented reality app usage on customer relationships

and continuance intention: The role of shared social

experience, Journal of Business Research. Volume 166,

2023, 114092, ISSN 0148-2963,

https://doi.org/10.1016/j.jbusres.2023.114092"

Clare Sullivan. EU GDPR or APEC CBPR? A comparative

analysis of the approach of the EU and APEC to cross

border data transfers and protection of personal data in

the IoT era, Computer Law & Security Review, Volume

35, Issue 4, 2019, Pages 380-397, ISSN 0267-3649.

https://doi.org/10.1016/j.clsr.2019.05.004.

Leveraging AI to Mitigate Risks in Yoga Practice: A Real-Time Posture Correction Application

249

Bendik Bygstad, Egil Øvrelid. Towards digital agility: the

best practices for e-health. Procedia Computer Science,

Volume 239, 2024, Pages 2022-2029, ISSN 1877-0509.

https://doi.org/10.1016/j.procs.2024.06.388.

Naveen G. Halappa, Integration of yoga within exercise and

sports science as a preventive and management strategy

for musculoskeletal injuries/disorders and mental

disorders – A review of the literature, Journal of

Bodywork and Movement Therapies, Volume 34, 2023,

Pages 34-40, ISSN 1360-8592.

https://doi.org/10.1016/j. jbmt.2023.04.055.

icSPORTS 2024 - 12th International Conference on Sport Sciences Research and Technology Support

250