Community Research and Development Information Service - CORDIS

Periodic Report Summary 1 - REVAMMAD (Retinal Vascular Modeling, Measurement and Diagnosis)

The REtinal VAscular Modeling, Measurement And Diagnosis (REVAMMAD) project ( is training a new generation of scientists able to effectively translate the latest vascular modeling theory and computerized image analysis techniques into effective interventions for some of the most important chronic medic conditions afflicting the EU, including hypertension and diabetes. It will particularly ensure that there is rich clinical and industrial involvement to ensure that the training is focused with end-users and exploitation in mind.

The vasculature undergoes changes in response to early stages of these diseases, reflecting fundamental physiological processes within the vessels. The retina provides a unique “window onto the vasculature,” allowing it to be viewed and measurements made in vivo, and advances in imaging technologies make it increasingly possible to measure subtle changes using computer vision algorithms, including through routine medical checks such as eye tests.

The objectives of the research programme are:

- To provide Early Stage Researchers with a deep understanding of the vasculature, its structure and properties, its flow dynamics, and of methods to model and measure it;
- To advance fundamental theory/knowledge via vascular models validates using empirical data, as a basis for research into physiology and disease;
- To utilize these to develop new techniques leading to effective clinical interventions using modeling, measurement and clinical understanding;
- To ensure the clinical and commercial relevance of all REVAMMAD research projects and to generate real impacts from the project;
- To construct a data warehouse for data and software components allowing comparative studies;
- To combine the above to develop an interoperable toolset for vascular modeling and analysis.

Progress against initial objectives is evident. REVAMMAD has exceeded expectations in the delivery of a high quality 'ad hoc' training programme for our fellows, with four out of the five initially planned workshops delivered within the first two years of the project. This high level of training has been combined with great efforts to spark collaborations between fellows and partners. We are delighted to report that collaboration have commenced between modellers and image processes to build flow models of real vessel networks, improved methods for extracting measurements given the topology of the retina (shape analysis, super-resolution imaging and montaging, tortuosity) and characterize their relationship to disease, developed modelling methods for retinal vessels and implemented an environment for shared experimental analysis.

Progress is tangible in all strands of the project. A new fluid-structure interaction model for retinal vasculature is now in place and biological experiments on occlusion in chicken chroioallantoic membrane have started. It is expected that these experiments will provide data for modelling work in the second part of the project. Also, fellows are deploying new methods using laser speckle imaging to obtain flow rates from the retina to detect vasomotion.

One of our fellows has prepared an initial network joining algorithm and will next liaise with modellers to translate the empirical network to model. This work will complement the in depth study of tortuosity in FFA and corneal images conducted in the first two years of the project. Also, there have been considerable advances in the algorithms for segmentation of conjuctiva vessels and tracing of corneal nerves and the integration of optic nerve head and vessel segmentation to calculate derived features. Furthermore, the initial work for a health economic model for diabetic retinopathy screening programme is now in place.

Models have also been generated to detect measurable changes in retinal calibre in progression of diabetic retinopathy and under exercise regime and mouse models are studying the microglial role in vascular development under normoxic and hypoxic conditions.

Initial algorithm work and data set development for the identification of area concentration of lesions in DR development have been achieved within the first two years of the project together with a super-resolution algorthm to improve enhancement of retinal images and 3D model registration.

In addition, a prototype of Data Warehouse is now complete and testing has beginning, facilitating crucial work in the benchmarking of algorithms. This has been combined with initial work on deep ensemble learning. The Data Warehouse is a major achievement for REVAMMAD given that it will allow for the incorporation of datasets and computations which will support the management of retinal images.

In summary, REVAMMAD is on track to 'produce' a new generation of scientists with expert knowledge that spans from the academic to the clinical and industrial sectors.

In addition to this intangible benefit, REVAMMAD is on track to pursue IP exploitation routes in the second part of the project, and especially, towards the end of the project. It is still expected that IP exploitation will be possible for the Data Warehouse and package modelling software components. It is expected that these tools will have great impact in the diagnosis, prognosis and prevention of diseases such as diabetes, hypertension, stroke and coronary heart disease and retinal diseases.

Project Website:
Project Outreach blog (maintained by ESRs):

Contact details:
REVAMMAD Coordinator: Professor Andrew Hunter, Pro Vice Chancellor and Head of the College of Science, University of Lincoln
REVAMMAD Project Manager: Pilar Pousada Solino, University of Lincoln ( 01522 835087

Photo 1: Detection of lesions
Photo 2: Found and missed candidates from ring contrast detector & Found and missed candidates from template matching + ring contracts detector


Carolyn Williams, (Research Manager)
Tel.: +441522886642


Life Sciences
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top