Work Package 5

DECODE cloud platform design and integration

WP5 - DECODE cloud platform design and integration [Months: 3-48]


Description of Work and Role ofSpecific Beneficiaries/Partner Organisations: The design, development and integration of the DECODE cloud platform will be undertaken jointly by the partners UOI, BIOIRC, MedApp, AGH UST. UOI will be the lead beneficiary.

Task 5.1: User requirements, specifications and architecture of the DECODE platform (UOI, BIOIRC, MedApp, AGH UST, M3-M15) This task deals with the collection of the user requirements, the functional and non-functional specifications and definition of the conceptual design of the platform. The user requirements will be collected from the beginning of the project, in terms of functionalities, ease of use and interfaces. From the operative point of view, the objective is to immediately provide the end users with valuable information and efficiently support in evaluating and optimizing the DCB design process. Additionally, the overall and the detailed DECODE platform architecture will be determined. The main modules of the DECODE platform and the main functions of each module will be identified and analyzed, as well as the technical specifications, the communication and the dependencies. The consortium members, according to their expertise (model developers, platform developers, DCB Industry experts, vascular surgeons), will discuss and analyze the needs, constraints and requirements of the system aiming to determine the key parameters which need to be addressed. Lead partner: UOI. Associated deliverables: D5.1

Task 5.2: Cloud-based framework, Data and Models Repository (UOI, BIOIRC, MedApp, AGH UST , M3-M48) This task deals with the establishment of a cloud-based framework consisting of a frontend accepting the connections from the clients and connecting them to the backend(s). More specifically, the frontend will be an open source web server (e.g. NginX) allowing for load balancing over multiple frontends and backends. Concerning the backend, it will be a custom developed set of application(s) based on a powerful open-source web application framework (e.g. Django). These application(s) will enalbe to authenticate a user, upload and download data from the repository, execute converter commands on the stored data, give access to stored 3D WebGL models and provide help functions and documentation to the user. The backends will connect to the DECODE database, where the authentication and application configuration will be stored. The database server will be an established open source product (e.g. PostgreSQL). The data and models repository will enable secured access storage of model data, in-silico results and program code (converters, renderers). It will provide data security (RAID) and backups. Finally, a different backup and securing model will be used for privacy related data, including removal option. To this end the following sub-tasks will be considered: (i) analysis of the databases requirements, (ii) configuration of the databases, (iii) development of the backends, (iv) deployment to data center, (v) testing, (vi) monitoring, and (vii) maintenance.

Lead partner: UOI. Associated deliverable: D5.2

Task 5.3: Model data converters and integration (BIOIRC, UOI, MedApp, AGH UST, M3-M48) The establishment of different modules is a prerequisite for the in-silico study as described in Task 4.2 and Task 4.3. Converter programs will be developed to convert the model resulting from one simulation to the required input format for the next stage. Converter programs are also necessary to merge models from different levels into the integrated model. Another type of converters will enable the upload in non-resident format to be converted to a resident format, and to convert resident format to user requested format on download. Converter programs will be stored in the data and models repository for online use during upload/download. The users will be able to upload their own converters, download the convertersto include them in their own toolchain, or to provide new functionality. To this end the following sub-tasks will be considered: (i) identification and specification of the required resident data formats and models, (ii) development, testing and maintenance of the required converter programs.

Lead partner: BIOIRC. Associated deliverable: D5.3

Task 5.4: High Visual representation tool (UOI, BIOIRC, MedApp, AGH UST, M3-M48) This task is based on OpenGL and WebGL technology. More specifically, OpenGL is a cross-platform technology allowing the rendering of 3D models and interacting with them. WebGL is a technology allowing OpenGL models to be rendered and interacted with inside a browser. Thus models can be downloaded over the net and locally viewed and used in the client browser. This implies that DECODE Cloud platform will be able to be used in parallel users, such as DEB biomedical companies, as well clinical centers that could perform in parallel clinical trials. Specific OpenGL shader code may be needed to provide all required functionality and to make appropriate use of the OpenGL acceleration in current graphics hardware. The Renderer programs will be stored in the data repository for online remote rendering. Users will be able to upload their own renderers providing alternative views or download them to include in their own toolchain. This includes the development and evaluation of the required renderer programs and OpenGL shader code. Concerning WebGL, the model must be programmed in JavaScript, the native browser extension language. Basic WebGL frameworks exist which enable the concentration of the developer on the application itself. One or more generic renderer programs will be developed which can display the models and simulation results at the different stages. User interaction will be provided to rotate and zoom the model, to selectively suppress or include parts of the model and/or results, and to change colors and transparency. Finally, model and in-silico data will be imported in the generic renderer as a set of data files in a standardized format.

Lead partner: UOI. Associated deliverable: D5.4

Task 5.5: Accessible interaction – VR and AR visualizations (MedApp, BIOIRC, UOI, AGH UST, M3-M48) Holographic visualization ofPAD data will be performed using HoloLens headsets produced by Microsoft. An additional extension to CarnaLife Holo application (MedApp S.A.) will be developed. The new application version will contain options dedicated to PAD data processing including filtering, volume rendering, segmentation and measurements options. Application will be working with every kind of volumetric DICOM data. Lead partner: MedApp. Associated deliverables: D5.5