Markerless Motion Capture for 3D Clinical Gait Analysis (IMAGine) (IMAGine)

Is Markerless Three-dimensional Motion Analysis a Viable Alternative to Traditional Gait Analysis? (IMAGine Study)

Many children have trouble with walking and frequently trip and fall. To understand the extent of the changes in their walking, but also to inform and evaluate possible interventions to improve their walking, a gait analysis is carried out. Gait analysis is the assessment of how someone walks. It requires the accurate placement of markers on the skin in anatomical locations. The positions of these markers are then measured using infrared cameras and the results are displayed as movement graphs. These graphs are used in clinical decision making. However, this assessment can be time-consuming and uncomfortable for the child, and it may trigger anxiety, leading to an unnatural walking pattern. This style of assessment may also not be suitable for some children who are either very young or have additional sensory needs.

One potential solution to these problems is to use motion capture systems that do not require markers. Instead, these systems use high-resolution video cameras to capture human motion. Artificial intelligence is then used to identify human body features from the video footage and produces movement graphs. Known as markerless motion capture, this emerging approach has been developed over the last few years and has demonstrated promising results compared to the current marker-based method in adults and children.

However, it is not known how well markerless motion capture works for children with gait abnormalities, which is what this study aims to find out. Therefore, fifty children with movement difficulties will walk barefoot over a walkway with and without markers attached, while markerless and marker-based cameras record all trials simultaneously.

This study will then compare the movement graphs created by the markerless and marker-based motion capture systems. The results of this study will also enable Alder Hey to be the lead clinical gait lab user of markerless technology in the UK, helping other gait labs to adopt the technology.

Study Overview

• Status: Completed
• Conditions: Paediatric Patients

Detailed Description

Background information and rationale Many children have impaired gait patterns. This can lead to an increased risk of tripping and falling, increase anxiety and ultimately reduce participation in physical activity with peers. Early identification of gait pathology requires a precise and accurate assessment of the gait, identifying the subtleties and how and to what extent they deviate from a typically developing gait.

The importance of gait analysis in children and adults has been widely demonstrated to provide crucial information for determining the extent of functional gait problems due to pathology and for their follow-up over time. It can also help to assess and plan surgical and rehabilitative interventions aimed at reducing functional limitations.

Clinical gait analysis is also used to inform and evaluate the outcomes of clinical interventions, such as orthopaedic or neurosurgical procedures, and to understand abnormalities, particularly in undiagnosed conditions.

Three-dimensional marker-based gait analysis provides comprehensive data on normal and pathological gait, which is crucial in clinical practice as it provides objective information on joint motion (kinematics), spatio-temporal data such as walking speed, step and stride length, and joint moments and power (kinetics).

The current gold standard in gait analysis requires the placement of small adhesive markers on the skin in precise anatomical locations. The 3D positions of these markers are measured using infrared cameras to identify and track the markers throughout the assessed gait. The results are then presented as joint angle and moment diagrams.

There are several problems with this approach for both the clinician and the patient. Firstly, it is a time-consuming approach, taking approximately 2 hours per patient. Secondly, it requires the patient to wear minimal clothing, which can be uncomfortable for some children and, coupled with the awareness that they are being observed, can lead to feelings of anxiety about the procedure and cause unnatural walking patterns. Clinicians are also affected by the time-intensive nature of the procedure, which limits the number of patients they can see over the course of a week. They can often see high variability in the children's gait due to the controlled laboratory environment. The application of the markers carries a risk of error due to incorrect marker placement, particularly if the child has bony deformities. This includes a systematic bias in marker placement by practitioners, which can make consolidation of data sets and multi-centre collaborations difficult. Furthermore, regardless of a highly skilled practitioner and accurate marker placement, it does not account for the movement of soft tissue artefacts such as skin, muscle and tissue covering the bone.

A potential solution to these problems comes from the development of markerless motion capture systems. Theia3D is a markerless motion capture alternative to the traditional marker-based approach that does not rely on skin-based markers to capture 3-dimensional human motion. Theia uses high-resolution 2-dimensional video cameras to capture videos of human motion and uses a machine learning algorithm-based approach to identify specific features of the human body within the videos. The neural network for feature identification has been trained on over one million images of humans. Deconvolution layers are used to generate spatial probability densities for each image, representing the likelihood that an anatomical feature is in a particular location. Theia targets salient features of the human body, rather than markers that are attached to the skin as close as possible to the specific anatomical landmark.

Theia3D has been shown to provide highly comparable results to the current marker-based approach for the hip, knee and ankle joints of adults during walking. Theia3D has also been shown to work effectively with normal clothing. However, this technology has not been studied in children with gait abnormalities. This is vital to establish whether it can be used effectively in children, as there is huge potential to improve the experience of patients and clinicians during clinical gait analysis. As Theia3D does not rely on markers attached to the skin, the problems caused by markers are reduced or eliminated. For example, markers can be occluded when children require walking aids (crutches, frames, etc.), neurodivergent children often struggle with markers adhering to their whole body, marker removal can be painful, markers are often knocked off requiring replacement and recalibration, and sweating can make marker adherence difficult. A markerless approach also has the potential to reduce overall data collection time immensely and eliminate errors caused by inaccurate marker placement.

The Theia3D markerless approach has huge upside potential. The robustness and accuracy of the system in tracking children with pathology needs to be tested to explore the potential use. This potential includes increased patient data collection, ease of collection and a more convenient experience for both patient and clinician. Therefore, this study aims to assess the accuracy and interchangeability of the markerless system.

Research objectives Could markerless motion analysis technology (Theia3D) be used as an alternative to the routine marker-based approach in clinical gait analysis for children with movement impairments?

Objectives:

Therefore, the primary objectives are:

1. to establish if Theia3D and the existing marker-based approach can be used interchangeably to determine lower body joint angles, moments, and powers during overground walking in children
2. to determine the feasibility of theia3D by comparing the presence or absence of clinically important features based on reports derived separately from marker-based and markerless data.

Secondary objectives:

1. Is markerless motion capture capable of identifying gait abnormalities in children (e.g. outtoeing, intoeing, toe walking, crouch gait)?
2. Do clinicians interpretation of gait analysis results differ between markerless and marker-based recordings?
3. Does the current marker-based approach cause children to have unnatural walking patterns?
4. Can Theia 3D identify lower body joint angles, moments and powers during overground walking in children when the legs and thorax is covered by clothing?

Methodology

This study will not require any changes to the participants experience during clinical gait analysis. The Theia markerless system will work simultaneously alongside the current routine. There are no further risks associated with the study as there are no additional tasks/ procedures carried out for this study. The current routine clinical gait analysis procedure as part of the patients normal care is detailed below:

Documents and Forms:

A full medical history will be taken by an experienced physiotherapist/ clinical scientist. The patient/ legal guardian will be asked to fill in the Gillette Functional Assessment Questionnaire as a self-assessment. Consent forms for the gait analysis and an additional informed consent and assent form regarding the markerless research study will be completed by parents.

Physical Examination:

A standardised physical examination is carried out on every patient that is seen for gait analysis. The process of the physical examination is explained to patient and family and their consent taken. A physical examination on an examination couch is carried out by two appropriately trained and experienced physiotherapist/clinical scientist with the assistance of one engineer. Measurements/ assessments include joint ranges, muscle strength, muscle tone and dynamic spasticity, any spinal curvatures and foot deformities described.

Video (2D) gait analysis:

Separate walking trials with the patient walking up and down for 3 walks over a 10m walkway will be collected with different clothing/ barefoot/ with shoes/ AFOs/ insoles, based on the clinical indication for the patient. Walking trials take place without any markers placed on the patient. 10 markerless Theia 3D cameras, including 2 cameras that are positioned in front of and to the side of the will be used to capture the walking conditions.

Marker-based and markerless (3D) gait analysis:

Software & equipment necessary for this data collection includes:

• 12 marker-based 3D infra-red cameras
• 4 force platforms integrated into the laboratory floor
• 10 markerless Theia 3D cameras This equipment is currently all expertly positioned within the Gait Laboratory at Alder Hey Children's Hospital/ North West Movement analysis Centre (NWMAC). The markerless and marker-based cameras will collect one static standing trial and at least 5 walking trials at self-selected speed over a 10m walkway. The marker-based and markerless cameras capture data synchronously, so no additional walking trials are required to the current protocol.

Participants Participant eligibility will be confirmed by an experienced physiotherapist/ clinical scientist working in the gait Laboratory at Alder Hey Children's Hospital/ North West Movement analysis Centre (NWMAC).

Informed consent All patients that are booked at the clinical gait analysis service for a 3D assessment will be deemed suitable and will be invited to take part in this study.

The patients will also be made aware of the study and will be informed that they will receive an information sheet in the post during the phone call when the gait lab assessment appointment is booked.

The patients will receive the study information sheet 1-3 weeks before their clinical gait assessment. The study will be briefly explained to them on arrival and the parents/ legal guardians and children will have time to ask all their questions. They will then be asked if they want to take part and if so, they will be asked to fill in the informed consents on the arrival date. If they do not want to partake or if the participant is unable to give consent it will have no impact on their gait analysis and they will not be included in this study.

The clinician who is carrying out the gait assessment with the child will obtain the informed consent from the person with parental responsibility for the child. All clinicians have been trained to obtain the informed consent and are obtaining informed consents on a routine basis for the clinical gait analysis.

All clinicians who will obtain the informed consent will be trained in the research study and will be able to answer all relevant questions. The children will also be offered to fill in an assent form.

Participants can withdraw by telling the clinician or any member of staff in the gait lab during data collection. Alternatively, they can withdraw at any time after the data collection by emailing the Chief investigator Dr Henrike Greaves at Henrike.greaves@alderhey.nhs.uk. If participants do not want to take part in this study anymore then the investigators will stop recording using the markerless system and delete any data that may have been collected. If the participant decide this retrospectively, then the investigators will delete the data that was collected and remove any processed data that may have been generated.

Arrangements for participants and their legal guardians who might not adequately understand English:

Patients will have received study information that will inform them about the process that includes the markerless cameras in appropriate language for their age or level of understanding. The parents/ legal guardians will also receive a study information sheet exclusively in English. However, certified interpreters are booked for appointments with patients and their parents/ legal guardians who don't adequately understand English to translate the information to them. The interpreter will translate all relevant information, as well as the information letter and informed consent to the participant.

Data storage and analysis All biomechanical gait data from the markerless and marker-based gait assessments will be securely stored on the Alder Hey server in the gait laboratory at Alder Hey Children's NHS Foundation Trust. The data analysis is carried out by the research team at the gait laboratory at Alder Hey Children's NHS Foundation Trust. The markerless video technology uses 2D video and artificial intelligence to extract the detailed information about the child's gait. Only clinicians within the gait laboratory, Joel Kearney and his academic supervisors will have access to those videos. The videos will not leave the gait lab. In the process of the data analysis at Alder Hey the video data will be converted into joint movement (kinematics), spatio-temporal data (walking speed, step and stride length, cadence, etc) and joint moments and powers. Thereby the participant data (e.g. video data) becomes represented as numerical biomechanical data which can then be anonymised. This anonymised data will then be shared with the research team at Liverpool John Moores University (LJMU) and a statistician for further analysis.

Alder Hey Children's NHS Foundation Trust is the sponsor of this research and is responsible for looking after this information.

Data storage The clinical data which is part of the routine gait analysis will be stored clinically for the patients records indefinitely at Alder Hey Children's NHS Foundation Trust. The markerless video data which is part of the research study will be stored for 5 years after the completion of the study at Alder Hey Children's NHS Foundation Trust. The processed and anonymised markerless data will be stored on a password-protected external hard drive for 5 years after the completion of the study at LJMU. This password protected external hard drive can only be accessed by the researchers within this study and will be stored in a locked cabinet at LJMU.

Personal data:

Only members of the medical care team at Alder Hey will have access to the participants personal data as part of the routine gait analysis. This information will not be accessed by or shared with the research team at Liverpool John Moores University. The signed informed consent form will be locked in a cabinet at LJMU and only the researchers involved in this study have access to it.

Participant confidentiality Personal data during the trial will be stored and used in accordance with Alder Hey's Standard Operating Procedure for confidentiality, protection, and breach of personal data in relation to research subjects. This ensures that all personal data collected during the trial is recorded, handled, and stored in such a way that is satisfies the requirements of the UK General Data Protection Regulation and requires data to be anonymised as soon as it is practical to do so.

All electronic patient-identifiable information will be held on a secure, password-protected database accessible only to authorised personnel within the research team and the gait laboratory. Paper forms with patient-identifiable information will be held in secure, locked filing cabinets within a restricted area at LJMU. The processing of the personal data of participants will be minimised by the use of a unique participant trial number. Personal data on all documents will be regarded as confidential.

The Investigator site will maintain the patient's anonymity in all communications and reports related to the research. Data Breaches will be highlighted to the relevant site staff and reported as required by the UK GDPR and Data Protection Act 2018.

Statistical analysis To establish face validity, markerless-derived joint angles, moments, and powers will be compared to the traditional marker-based approach using statistical waveform comparison techniques (e.g. Linear Fit Method, Statistical Parametric Mapping, Generalised Additive Mixed Models) and root mean square differences. Generalised Additive Mixed Models (GAMM) will be fitted to the data to identify the systematic and random effects of measurements by device type, time, age, subject, and trial. Spline expansions will be used to model the observed nonlinearities in the measurements. Approximate F test statistics will be obtained to test the null hypothesis of equality of the measurements across the device type, controlling for age, and repeated measurements within subjects.

Sample size/ power calculations The target sample size for the trial is 100 participants. This will allow detection of a clinically meaningful moderate standardised effect size of 0.33 with a two-sided 5% significance level, 90% power. The standardised effect size of 0.33 corresponds to a difference of 5 degrees in joint angles, with a standard deviation of 15 degrees. A difference of 5 degrees is considered a minimal detectable change in joint kinematics during walking to report a significant change (Wilken et al., 2012).

Outcomes The primary outcome is to establish if Theia3D and the existing marker-based approach can be used interchangeably to determine lower body and trunk joint angles, moments, and powers during overground walking in children.

The secondary objectives are to investigate whether:

1. Clinical experts' interpretation of gait analysis results differ between markerless and marker-based recordings?
2. Current marker-based approach cause children to have unnatural walking patterns?
3. Theia3D can identify lower body joint angles, moments and powers during overground walking in children when the legs and thorax is covered by clothing?

Ethical considerations

Declaration of Helsinki The Investigator will ensure that the trial is conducted in accordance with the principles of the Declaration of Helsinki.

Guidelines for Good Clinical Practice The Investigator will ensure that the trial is conducted in accordance with relevant regulations and with the principles of Good Clinical Practice.

Ethical conduct of the trial and ethical approvals The protocol, patient information sheet, informed consent form and any other information that will be presented to potential trial participants (e.g. advertisements or information that supports or supplements the informed consent process) will be reviewed and approved by an appropriately constituted, independent Research Ethics Committee (REC).

NHS Research Governance Once HRA & HCRW approval is in place for the trial, sites will confirm capability and capacity to participate in the trial.

Protocol amendments All amendments will be generated and managed according to the trial office standard operating procedures to ensure compliance with applicable regulation and other requirements. Written confirmation of all applicable REC and local approvals must be in place prior to implementation by Investigators as applicable for the amendment type.

Protocol Compliance and Deviations Protocol compliance is fundamental to GCP. Prospective, planned deviations or waivers to the protocol are not allowed. Changes to the approved protocol need prior approval unless for urgent safety reasons.

Trial reports This protocol will comply with all current applicable Research Ethics Committee and Sponsor reporting requirements.

End of Study Completion of data analysis for this study will be approximately 1 year after data collection starts. The end of the study is reached, and the trial will stop when the stated number of participants is reached.

The research data will be kept for 5 years after the work based on these data has been published and/or after the research project has ended. The markerless video data that is part of the research study will be stored for 5 years after the completion of the study at Alder Hey Children's NHS Foundation Trust. The anonymised data will be stored for 5 years after the completion of the study on a password protected hard-drive in a locked cabinet at Liverpool John Moores University (LJMU).

• Study Type: Observational
• Enrollment (Actual): 100

Contacts and Locations

Study Locations

• United Kingdom
  • Merseyside
    • Liverpool, Merseyside, United Kingdom, L14 5AB
      • Alder Hey Children's NHS Foundation Trust

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

Participant referred to the gait lab to carry out gait analysis

Description

Inclusion Criteria:

• Referred to the gait lab to carry out a 3D gait analysis
• Able to cope with the demands of the 3D gait analysis
• Able to comprehend and understand instruction in English in order to consent and safely undertake the clinical gait analysis
• Aged between 6-18 years of age

Exclusion Criteria:

• Non-ambulatory, unable to walk.
• Aged below 6 years of age
• Unable to comprehend and understand instruction in English in order to consent and safely undertake the clinical gait analysis

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort
Children with gait abnormalities

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
To establish if Theia3D and the existing marker-based approach can be used interchangeably to determine lower body and trunk joint angles during overground walking in children.	Markerless-derived joint angles (measured in degrees) will be compared to the traditional marker-based approach using statistical waveform comparison techniques.	Baseline
To establish if Theia3D and the existing marker-based approach can be used interchangeably to determine lower body joint moments during overground walking in children.	Markerless-derived joint moments (Nm) will be compared to the traditional marker-based approach using statistical waveform comparison techniques.	Baseline
To establish if Theia3D and the existing marker-based approach can be used interchangeably to determine temporo-spatial parameters during overground walking in children.	Markerless-derived temporo-spatial parameters will be compared to the traditional marker-based approach using root mean square differences and paired t-tests.	Baseline

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Do clinical experts' interpretation of gait analysis results differ between markerless and marker-based recordings?	Two clinical experts at Alder Hey will review a clinical report consisting of the 2D videos, physical examination, foot pictures, and 3D gait curves (joint kinematics, kinetics, powers and ground reaction forces). The difference between the two reports will be that one report contains the 3D markerless gait curves derived from Theia3D and the other one will contain the marker-based 3D gait data. The clinical experts will be blinded and randomly assigned to independently review the gait report. Clinical experts are instructed to review the full gait report and formulate the most likely impairments affecting the gait pattern, using Richard Baker's impaired focused approach. Reporting will follow the mandatory data interpretation and reporting guidelines set out by CMAS. The results will then be compared to assess whether the gait reports were interpreted differently and resulted in different impairments when reviewed.	Baseline
Does marker-application lead to different joint kinematics potentially leading to unnatural walking patterns in children?	Children will be instructed to walk barefoot without any marker application. The children will be asked to wear either 2-piece swimsuit (e.g. bikini/tankini) or gym type/lycra shorts and vest top for females and swimming trunks/briefs for males. The walk will be collected with 10 markerless cameras and analysed using Theia3D. The gait data will then be compared to the gait collected with the application of markers and assessed for differences. Markerless-derived joint angles (measured in degrees) will be compared to the markerless gait trials with and without marker application	Baseline
Does marker-application lead to different temporospatial parameters potentially leading to unnatural walking patterns in children?	Children will be instructed to walk barefoot without any marker application. The children will be asked to wear either 2-piece swimsuit (e.g. bikini/tankini) or gym type/lycra shorts and vest top for females and swimming trunks/briefs for males. The walk will be collected with 10 markerless cameras and analysed using Theia3D. The gait data will then be compared to the gait collected with the application of markers and assessed for differences. Markerless-derived temporo-spatial parameters will be compared to the markerless gait trials with and without marker application using root mean square differences and paired t-tests.	Baseline
Can Theia3D identify lower body joint angles during overground walking in children when the legs and thorax is covered by clothing?	Children will be instructed to walk barefoot wearing their normal clothing that they have arrived with to the gait lab. The walk will be collected with 10 markerless cameras and analysed using Theia3D. The markerless gait data will then be compared to the gait collected to the gait wearing either 2-piece swimsuit (e.g. bikini/tankini) or gym type/lycra shorts and vest top for females and swimming trunks/briefs for males. The comparison between walking with normal and less clothing will be carried out for joint angles of the thorax, pelvis, hip, knee, and ankle using a statistical waveform comparison technique. This analysis will help to identify whether the gait data with clothing led to significantly different results compared to walking with less clothing.	Baseline
Can Theia3D identify joint moments during overground walking in children when the legs and thorax is covered by clothing?	Children will be instructed to walk barefoot wearing their normal clothing that they have arrived with to the gait lab. The walk will be collected with 10 markerless cameras and analysed using Theia3D. The markerless gait data will then be compared to the gait collected to the gait wearing either 2-piece swimsuit (e.g. bikini/tankini) or gym type/lycra shorts and vest top for females and swimming trunks/briefs for males. The comparison between walking with normal and less clothing will be carried out for lower body joint moments (Nm) using a statistical waveform comparison technique. This analysis will help to identify whether the gait data with clothing led to significantly different results compared to walking with less clothing.	Baseline

Collaborators and Investigators

• Sponsor: Alder Hey Children's NHS Foundation Trust
• Collaborators: Liverpool John Moores University
• Investigators: Principal Investigator: Richard J Foster, BSc (hons), MSc, PhD, Liverpool John Moores University; Principal Investigator: Joel Kearney, BSc, MSc, Liverpool John Moores University

Publications and helpful links

General Publications

• Stewart C, Eve L, Durham S, Holmes G, Stebbins J, Harrington M, Corbett M, Kiernan D, Kidgell V, Jarvis S, Daly C, Noble J. Clinical Movement Analysis Society - UK and Ireland: Clinical Movement Analysis Standards. Gait Posture. 2023 Sep;106:86-94. doi: 10.1016/j.gaitpost.2023.08.006. Epub 2023 Aug 18. No abstract available.
• Wilken JM, Rodriguez KM, Brawner M, Darter BJ. Reliability and Minimal Detectible Change values for gait kinematics and kinetics in healthy adults. Gait Posture. 2012 Feb;35(2):301-7. doi: 10.1016/j.gaitpost.2011.09.105. Epub 2011 Oct 29.
• Kanko RM, Laende EK, Davis EM, Selbie WS, Deluzio KJ. Concurrent assessment of gait kinematics using marker-based and markerless motion capture. J Biomech. 2021 Oct 11;127:110665. doi: 10.1016/j.jbiomech.2021.110665. Epub 2021 Aug 3.
• McGinley JL, Baker R, Wolfe R, Morris ME. The reliability of three-dimensional kinematic gait measurements: a systematic review. Gait Posture. 2009 Apr;29(3):360-9. doi: 10.1016/j.gaitpost.2008.09.003. Epub 2008 Nov 13.
• Hallemans A, Van de Walle P, Wyers L, Verheyen K, Schoonjans AS, Desloovere K, Ceulemans B. Clinical usefulness and challenges of instrumented motion analysis in patients with intellectual disabilities. Gait Posture. 2019 Jun;71:105-115. doi: 10.1016/j.gaitpost.2019.04.016. Epub 2019 Apr 22.
• Hulleck AA, Menoth Mohan D, Abdallah N, El Rich M, Khalaf K. Present and future of gait assessment in clinical practice: Towards the application of novel trends and technologies. Front Med Technol. 2022 Dec 16;4:901331. doi: 10.3389/fmedt.2022.901331. eCollection 2022.
• Cimolin V, Galli M. Summary measures for clinical gait analysis: a literature review. Gait Posture. 2014 Apr;39(4):1005-10. doi: 10.1016/j.gaitpost.2014.02.001. Epub 2014 Feb 7.

Study record dates

Study Major Dates

• Study Start (Actual): March 31, 2025
• Primary Completion (Actual): December 22, 2025
• Study Completion (Actual): December 22, 2025

Study Registration Dates

• First Submitted: July 7, 2025
• First Submitted That Met QC Criteria: April 30, 2026
• First Posted (Actual): May 7, 2026

Study Record Updates

• Last Update Posted (Actual): May 7, 2026
• Last Update Submitted That Met QC Criteria: April 30, 2026
• Last Verified: July 1, 2025

More Information

Terms related to this study

• Keywords: children; Gait analysis; paediatric; markerless movement analysis; Theia3D; movement abnormalities
• Additional Relevant MeSH Terms: Central Nervous System Diseases; Nervous System Diseases; Movement Disorders
• Other Study ID Numbers: 334786 (Other Identifier: Integrated Research Application System)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No