Welcome to the 3rd DeepHealth newsletter! Throughout these communications, we will be providing more relevant news about the project in the last months.
A new year has started, and we hope that it will be better in many ways. Although the pandemic is not defeated, we keep working with enthusiasm, collaborating virtually inside the projects and participating in virtual events to spread knowledge of DeepHealth in the scientific community. We are also extending our project activity to study and investigate methods for analysing and fight COVID-19. Navigate through this newsletter to get a glimpse of our highlights during the last months!
Main Goals and Results
As announced in the previous newsletter, the first versions of the two libraries developed within the DeepHealth project, EDDL (European Distributed Deep Learning Library) and ECVL (European Computer Vision Library) are publicly available in GitHub. EDDL and ECVL are two of the major outcomes of the DeepHealth project, forming together with the associated front-end and back-end application, the DeepHealth toolkit. We are currently working to provide a stable integration of all the components to make the toolkit publicly available soon. We are also working on extending the libraries functionality, e.g. adding new models in EDDLL, and improving their performance. To this aim we organised internal Collaborative Workshops to experiment all together on selected project use cases, e.g. Skin Lesion Detection and Classification, using different models provided by the libraries and running on the highly heterogeneous HPC and cloud computing infrastructures provided by the project partners. As we obtained very valuable feedback from this activity, we are organising new workshops in the next months, and thanks to the results obtained so far, a new major release of the DeepHealth libraries will be released in the next weeks.
Great presence of DeepHealth at HiPEAC 2021! DeepHealth was invited to participate in two workshops on the 19th of January: in HeLP-DC, contributing with a general introduction of the project and a technical presentation about leveraging HPC and cloud infrastructures in DeepHealth, and in PARMA-DITAM, with an invited talk on the HPC applications cloudification.
On 12th January 2021, Marco Grangetto and University of Turin team, participated in ICPR2020, presenting in the “Medical and Industrial Imaging” session the demo “Lung nodules segmentation in CT scans by DeepHealth toolkit”. The demo showed how a deep learning pipeline can be executed on the University of Turin's hybrid HPC-Cloud infrastructure OpenDeepHealth.
On 5th November, Iacopo Colonnelli (University of Torino) presented the integration of Jupyter Notebook, with StreamFlow, a component of the OpenDeepHealth infrastructure, at the annual workshop organised by the GARR (the ultra-broadband network dedicated to the Italian research) consortium.
On 17th September, Monica Caballero (Everis) gave an online keynote at the Spanish Supercomputing Network (RES) 14th Users Conference, showing how DeepHealth addresses addresses Deep Learning's convergence with HPC and how it exploits these technologies for the benefit of biomedical applications.
Between 3rd and 5th of November members of DeepHealth project took part in the flagship event of the European Big Data and Data-Driven AI Research and Innovation community organised by the BDVA and the European Commission. Jon Ander Gomez (UPV) also presented DeepHealth experience, in the session concerning "Project perspective on Big Data and AI architectural pipelines and benchmarks"
The Big Data Value Association (BDVA) White Paper on AI in Healthcare was officially presented at the EBDVF 2020. It includes the DeepHealth project as a successful use-case in the application of AI-based technologies to Boost Biomedical Applications for Health. Monica Caballero (Everis) and Amalia Ntemou (WINGS) participated in the official presentation as contributors.
Marco Aldinucci (University of Turin) on 20th November presented at the Web Marketing Festival preliminary results on the CLAIRE-COVID activity also showing how DeepHealth uses Artificial intelligence and high-performance computing for the diagnosis of Covid-19 pneumonia.
Marco Grangetto and his team in the University of Turin and Città della Salute, have been working to investigate the possibility to use widespread and simple chestX-ray (CXR) imaging for early screening COVID-19 patients. Early results are presented in the International Journal of Environmental Research and Public Health.
