ORGANISATION/COMPANYIMT Atlantique / LaTIM
RESEARCH FIELDComputer science › Database managementMedical sciences
RESEARCHER PROFILEFirst Stage Researcher (R1)Recognised Researcher (R2)Established Researcher (R3)Leading Researcher (R4)
APPLICATION DEADLINE16/06/2021 00:00 - Europe/Brussels
LOCATIONFrance › Brest
TYPE OF CONTRACTTemporary
The laboratory of medical information processing (LaTIM UMR 1101, Inserm) is opening a PhD position on longitudinal follow-up of multi-modal medical images using artificial intelligence.
Statistics from the International Diabetes Federation reveal that the global prevalence of diabetes in 2019 is around 9.3% (463 millions) of the world population and will rise to 10.9% by 2045 (700 millions). As a common and high-risk complication of diabetes, diabetic retinopathy (DR) is a leading cause of visual impairment and blindness worldwide . Although regular screening is crucial for preventing blindness, the expected increase in the number of patients with diabetes means that the burden of screening and follow-up represent a substantial challenge .
Standard retinal screening usually employs color fundus photography (CFP) for DR diagnosis by examining the presence of retinal lesions such as microaneurysms, hemorrhages or exudates. For the automatic assessment of DR evolution, emerging imaging modalities could allow a finer analysis. In particular, ultrawide-field CFP (UWF-CFP) gives useful information on the periphery of the retina, not seen on standard photography. Structural optical coherence tomography (OCT) that produces few microns resolution cross sectional imaging can be enriched with angiography (OCTA) to highlight retina vessels non-invasively. Deep learning (DL) applied to UWF-CFP and OCTA longitudinal images represent a promising perspective for DR management .
The PhD thesis will take place in the context of the ANR RHU EviRed project whose aim is to develop and validate an expert system guiding ophthalmologists to improve diagnosis, prediction of evolution and decision-making during DR follow-up.
The main objective is to exploit follow-up UWF-CFP and OCTA examinations to improve DR severity and progression assessment, by taking advantage of both past and current examinations.
In this context, we hypothesize that visual patterns predicting disease progression exist in images and can be extracted through deep learning. Moreover, analyzing follow-up examinations may improve automated grading performance with respect to single-image strategies .
Description of work
To jointly analyze follow-up examinations, consecutive images must be put in a common reference frame. Therefore, in addition to OCTA/UWF-CFP multimodal registration, a unimodal registration step will be included. This will allow capturing local changes inside the retina between consecutive examinations. Difference images will be used as additional features for severity assessment and disease progression. Smarter difference operators will have to be designed to avoid the influence of registration errors and illumination variations between examinations .
Depending on the investigated representations of multi-modal longitudinal data, appropriate CNN architectures  including Siamese or CNN+RNN  strategies will be applied to assess and then predict the DR progression. Finally, visualization techniques will be employed for the extraction and analysis of spatio-temporal predictive patterns in follow-up examinations, to elucidate the changes and early-warning signs that should be looked for in examination records.
 K. Ogurtsova, J. da Rocha Fernandes, Y. Huang, U. Linnenkamp, L. Guariguata, N. H. Cho, D. Cavan, J. Shaw, and L. Makaroff, “IDF diabetes atlas: Global estimates for the prevalence of diabetes for 2015and 2040”, Diabetes Research and Clinical Practice, vol. 128, pp. 40–50, 2017.
 A. Bora, S. Balasubramanian, B. Babenko, S. Virmani, S. Venugopalan, A. Mitani, G. de Oliveira Ma-rinho, J. Cuadros, P. Ruamviboonsuk, G. S. Corrado et al., “Predicting the risk of developing diabetic retinopathy using deep learning”, The Lancet Digital Health, vol. 3, no. 1, 2021.
 S. Stolte and R. Fang, “A survey on medical image analysis in diabetic retinopathy”, Medical Image Analysis, vol. 64, p. 101742, 2020.
 G. Quellec, K. Charrière, Y. Boudi, B. Cochener, and M. Lamard, “Deep image mining for diabetic retinopathy screening”, Medical Image Analysis, vol. 39, pp. 178–193, 2017.
 M. Gong, J. Zhao, J. Liu, Q. Miao, and L. Jiao, “Change detection in synthetic aperture radar images based on deep neural networks”, IEEE Transactions on Neural Networks and Learning Systems, vol. 27, no. 1, pp. 125–138, 2015.
 F. Arcadu, F. Benmansour, A. Maunz, J. Willis, Z. Haskova, and M. Prunotto, “Deep learning algorithm predicts diabetic retinopathy progression in individual patients”, NPJ Digital Medicine, vol. 2, no. 1, pp. 1–9, 2019.
 Y. Yan, P.-H. Conze, M. Lamard, G. Quellec, B. Cochener, and G. Coatrieux, “Towards improved breast mass detection using dual-view mammogram matching”, Medical Image Analysis, 2021.
 S. Perek, L. Ness, M. Amit, E. Barkan, and G. Amit, “Learning from longitudinal mammography studies”, in International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 712–720, 2019.
Funding category: Financement public/privé
ANR RHU EviRed
PHD title: Doctorat en analyse et traitement de l’information et des images médicales
PHD Country: France
The following skills are required: strong theoretical and practical knowledge in applied mathematics, image processing, machine/deep learning, Python programming, organizational skills, fluent English for reading/writing scientific articles, interest in the fields of health and artificial intelligence
Deadline for application: 11th June 2021.
EURAXESS offer ID: 636542
Posting organisation offer ID: 97974
The responsibility for the jobs published on this website, including the job description, lies entirely with the publishing institutions. The application is handled uniquely by the employer, who is also fully responsible for the recruitment and selection processes.
Please contact firstname.lastname@example.org if you wish to download all jobs in XML.