Automated Fetal and Neonatal Movement Assessment for Very Early Health Assessment

Lead Research Organisation: King's College London
Department Name: Imaging & Biomedical Engineering

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description The award has developed a new method which can estimate motor behaviour and pose in human infants completely non-invasively (using video). This is important as it enables automated analysis which is amenable to exploration with machine learning and can greatly reduce the workload associated with a human observer.
Using MRI, we have also studied patterns of functional connectivity in the sensorimotor cortex across the neonatal period. We found that across the preterm period, functional connections within this network are rapidly developing and becoming more specific - highlighting the key importance for this time for this emerging system.
Exploitation Route The developed software has been made available for download in a Github repository
Sectors Healthcare
URL http://www.perinatal-functional-imaging.co.uk/research/
 
Description Developing responsible neurotechnology for infants and children with neurodevelopmental conditions
Amount £1,048,588 (GBP)
Funding ID EP/W035154/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 05/2022 
End 05/2025
 
Description Collaboration for general movements in infants 
Organisation Norwegian University of Science and Technology (NTNU)
Country Norway 
Sector Academic/University 
PI Contribution Research collaboration to use automated methods (using computer vision and deep learning) for studying and classifying infant general movements.
Collaborator Contribution Partners are providing large amount of data and access to their methods for comparision with those developed as part of our EPSRC project.
Impact none as yet
Start Year 2021
 
Description Collaboration with Chalmers University Sweden 
Organisation Chalmers University of Technology
Country Sweden 
Sector Academic/University 
PI Contribution We are collecting 3D camera and electro-magnetic tracking data which is being used to measure infant movements. We are collecting the data.
Collaborator Contribution Collaboration with Silvia Muceli who is an expert in human motor control and signal processing. She has post-doctoral fellow and a Marie-Curie fellowship to work on similar work so has been providing expertise about the associated analysis
Impact This is a multi-discplinary collaboration with a bioengineering group. This work has only started recently so there are no specific outputs yet.
Start Year 2019
 
Title Software for 'Unsupervised Human Pose Estimation through Transforming Shape Templates" 
Description Human pose estimation is a major computer vision problem with applications ranging from augmented reality and video capture to surveillance and movement tracking. In the medical context, the latter may be an important biomarker for neurological impairments in infants. Whilst many methods exist, their application has been limited by the need for well annotated large datasets and the inability to generalize to humans of different shapes and body compositions, e.g. children and infants. In this paper we present a novel method for learning pose estimators for human adults and infants in an unsupervised fashion. We approach this as a learnable template matching problem facilitated by deep feature extractors. Human-interpretable landmarks are estimated by transforming a template consisting of predefined body parts that are characterized by 2D Gaussian distributions. Enforcing a connectivity prior guides our model to meaningful human shape representations. We demonstrate the effectiveness of our approach on two different datasets including adults and infants. 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact Software was presented at the prestigious CVPR conference 2021 and has initiated a collboration with researchers at the Norwegian Technical University. 
URL https://github.com/lschmidtke/shape_templates
 
Description Paper accepted for CVPR (Computer Vision and Pattern Recognition) 2021 conference 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact "Unsupervised Human Pose Estimation through Transforming Shape Templates; Luca Schmidtke, Simon Ellershaw, Athanasios Vlontzos, Bernhard Kainz, Anna Lukens, Tomoki Arichi" accepted for the prestigious 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). The paper presents a novel method for learning pose estimators for human adults and infants in an unsupervised fashion.
This will be presented later in the year at the meeting. This meeting reaches a wide audience including academic, industrial, and media outlets as it is the most important meeting in the field annually.
Year(s) Of Engagement Activity 2021
URL http://cvpr2021.thecvf.com/
 
Description Parent Power family engagement event 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Public/other audiences
Results and Impact I took part in a "Parent Power" event for families from Lambeth and Southwark organised by King's College London. This event was designed to provide a forum for families from the local community to hear about research and to be able to engage with researchers to ask questions. I spoke about my own work using MRI to study brain development and autism, and then had an open discussion and question/answer session with these families.
Year(s) Of Engagement Activity 2021
 
Description Perinatal Functional Imaging Twitter 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact Twitter page (Kings_PFIG) to transmit current views, advertise positions, describe and disseminate current work. Currently 352 followers
Year(s) Of Engagement Activity 2018,2019,2020,2021
URL https://twitter.com/Kings_PFIG
 
Description Perinatal Functional Imaging group website 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Academic Group website, developed to provide information about the research work and provide access for resources such as data, analysis tools, and links to publications. The website is updated regularly to reflect current work and provide contract details.
Year(s) Of Engagement Activity 2018,2019,2020,2021
URL http://www.perinatal-functional-imaging.co.uk
 
Description William Little Foundation - evaluation of need for early intervention and recognition in cerebral palsy 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Supporters
Results and Impact Inivited to participate in the working group and to review the draft document of an assessment and statement from the Willam Little Foundation about early recognition and intervention in cerebral palsy.
Year(s) Of Engagement Activity 2021