MARVEL
Assets
Identification of MARVEL building blocks
MARVEL Solution Assets
MARVEL is a Research and Innovation Action (RIA), so the expected Technology Readiness Level (TRL) is 5-6 on average, which involves a real-life environment technology validation and demonstration. This will be achieved, mostly, with the validation of the assets in the industrial-relevant smart city environments of Malta, Trento, and Novi Sad. Most of the assets will be tested in the MARVEL use cases.
The expected MARVEL framework will consist of 29 technologies developed under the project scope classified as:
1
Software/Hardware Developments
Τhis category refers to tangible outcomes, namely sensors and software solutions.
2
Methodologies
Τhey help to formalise user requirements into technological requirements.
3
Services
The word “Service” refers to the traditional meaning of IT services. These are future services offered around MARVEL, which aim at improving the effective use of MARVEL solutions and at providing in-depth customised assistance.
MARVEL Assets
the main aspects of the list of MARVEL Assets
- All
- Sensing and perception subsystem
- Data Management and Distribution Subsystem
- Audio visual and multimodal AI subsystem
- Optimised E2F2C processing and deployment subsystem
- E2F2C infrastructure
- User interactions and decision-making toolkit
- Security Privacy and Data Protection Subsystem
- software/hardware
- Methodology
- services
The Advanced MEMS microphones cater customer needs by gathering high quality, low noise acoustic data, which can be used for speech recognition, audio recording, and acquisition of surrounding noise for applications like noise cancellation.
- Type of component: Software/ Service
- Owner: IFAG
- Expected TRL: 8
- Licensing: Open source License TBD
SED@Edge is a solution to port state-of-the-art deep learning models for the detection of urban acoustic events on low-cost low-power microcontrollers.
- Type of component: Software
- Owner: FBK
- Expected TRL: 8
- Licensing: Open source License TBD
GRNEdge facilitates the synchronisation of different data streams (such as audio and video) for processing. It has the ambition to be a fully compliant plug-and-play sensor for smart city management This is possible once data-processing, such as audio-video anonymisation can also be done at the edge.
- Type of component: Hardware
- Owner: GRN
- Expected TRL: 6
- Licensing: Proprietary
The proposed system will enable target users to classify the crowd behaviour without breaking any privacy issues.
- Type of component: Hardware/Software
- Owner: UNS
- Expected TRL: 4-5
- Licensing: TBD
CATFLOW transforms road camera information to anonymous data. Through a few clicks, our dashboards then present this data and provide critical insights and detailed reports identifying a range of mobility types, trajectories, and junction turning ratios.
- Type of component: Software
- Owner: GRN
- Expected TRL: 7
- Licensing: Proprietary
EdgeSec aims to enable secure computing by focusing on the preservation of the confidentiality and integrity of applications and data, especially in the edge layer. Furthermore, EdgeSec transparently provides secure communication channels for data in transit to each participating node.
- Type of component: Software
- Owner: FORTH
- Expected TRL: 5
- Licensing: Proprietary
The Component VideoAnony will detect the faces of persons and perform anonymisation on the detected faces by blurring at the initial stage, and potentially with face conversion at a later stage.
- Type of component: Software
- Owner: FBK
- Expected TRL: 6
- Licensing: Open source License TBD
It helps anonymising audio data, possibly close to the source while preserving the acoustic content.
- Type of component: Software
- Owner: FBK
- Expected TRL: 4
- Licensing: Open source License TBD
devAIce SDK is a full-blown audio analysis SDK meant to enable customers to perform intelligent audio analytics on their local premises.
- Type of component: Software
- Owner: AUD
- Expected TRL: 6
- Licensing: Proprietary
The innovative aspects of a facilitating data management platform, such as DFB, lies mainly in the overall trustworthiness, performance, and quality of the implementation, focusing on security, scalability, redundancy or any other horizontal, non-functional need that is critical to the project at hand.
- Type of component: Software
- Owner: ITML
- Expected TRL: 6
- Licensing: Proprietary
The key capabilities and features offered by the Streamhandler platform are: Real-time monitoring and event-processing; Interoperability with all modern data storage technologies and popular data sources; Distributed messaging system; High fault-tolerance - Resiliency to node failures and support of automatic recovery; Elasticity - High scalability and Security (encryption, authentication, authorisation).
- Type of component: Software
- Owner: INTRA
- Expected TRL: 8
- Licensing: Proprietary
In MARVEL we expect to provide template data flows for typical smart cities solutions and functionalities for agent management in the edge.
- Type of component: Software
- Owner: ATOS
- Expected TRL: 6
- Licensing: Apache 2.0
HDD is a set of distributed adaptive data delivery and access algorithmic schemes for guaranteeing real-time delay requirements while effectively prolonging network lifetime in wireless edge networks. HDD targets large-scale network deployments which are characterised by an inherent computational intractability when it comes to the algorithmic management of the available data.
- Type of component: Methodology
- Owner: CNR
- Expected TRL: 6
- Licensing: MIT
It is a software deployed on all/some of the following: edge, fog, cloud. It includes a methodology improving performance for visual anomaly detection.
- Type of component: Software/ Methodology
- Owner: AU
- Expected TRL: 5
- Licensing: Open source License TBD
It is a software deployed on all/some of the following environments: edge, fog, cloud. It includes a methodology improving performance for audio-visual anomaly detection.
- Type of component: Software/ Methodology
- Owner: AU
- Expected TRL: 5
- Licensing: Open source License TBD
It is a software deployed on all/some of the following environments: edge, fog, cloud. It includes a methodology improving performance for visual crowd counting.
- Type of component: Software/ Methodology
- Owner: AU
- Expected TRL: 5
- Licensing: Open source License TBD
It is a software deployed on all/some of the following environments: edge, fog, cloud. It includes a methodology improving performance for audio-visual crowd counting detection.
- Type of component: Software/ Methodology
- Owner: AU
- Expected TRL: 5
- Licensing: Open source License TBD
This technology aims to improve upon current methods for automated audio captioning. It provides the audio analysis component that can be acted upon in the next stages of the MARVEL system.
- Type of component: Service
- Owner: TAU
- Expected TRL: 3
- Licensing: Open source License TBD
This technology aims to improve upon current methods for acoustic scene classification. It provides the audio analysis component that can be acted upon in the next stages of the MARVEL system.
- Type of component: Service
- Owner: TAU
- Expected TRL: 5
- Licensing: Open source License TBD
This technology aims to improve upon current methods for sound event detection. It provides the audio analysis component that can be acted upon in the next stages of the MARVEL system.
- Type of component: Service
- Owner: TAU
- Expected TRL: 5
- Licensing: Open source License TBD
This technology aims to improve upon current methods for sound event detection and localisation. It provides the audio analysis component that can be acted upon in the next stages of the MARVEL system.
- Type of component: Service
- Owner: TAU
- Expected TRL: 5
- Licensing: Open source License TBD
FORTH offers a real-time high-speed pattern matching engine that leverages the parallelism properties of GPGPUs to accelerate the process of string and/or regular expression matching. It is offered as a C API and allows developers to build applications that require pattern matching capabilities while simplifying the offloading and acceleration of the workload by exploiting the available GPU(s).
- Type of component: Software
- Owner: FORTH
- Expected TRL: 5
- Licensing: Proprietary
DynHP makes it possible to execute the interference and training steps of audio-video real-time analytics on devices with limited compute capabilities at the edge of the network, which otherwise cannot be used due to insufficient memory or energy or processing resources.
- Type of component: Methodology
- Owner: CNR
- Expected TRL: 4
- Licensing: MIT
Federated learning offers an effective technique for model training with privacy protection and network bottleneck mitigation at the same time. As a result, such distributed approach has been widely accepted as an effective technique for addressing the problem of large, complex, and time- consuming training procedures, and is also particularly suited for edge computing.
- Type of component: Software
- Owner: UNS
- Expected TRL: 5
- Licensing: Apache
This will be an evolution of FORTH’s open-source, Kubernetes-based execution environment for scientific data extraction and analysis, built upon an integration of Karvdash, Jupyter notebooks Argo Workflows. This execution platform already shows promise as an early realisation of the MARVEL E2F2C architectural framework, where the multimodal processing and AI/ML workloads relevant to MARVEL use-cases and pilots can be instantiated as orchestrated containers, and deployed via appropriate automation to execution sites selected by a dynamic online optimisation strategy.
- Type of component: Service
- Owner: FORTH
- Expected TRL: 5
- Licensing: Apache 2.0.
This component will help preparing of pilot and pre-production testbeds for the deployment of the MARVEL framework and its validation. In addition, it will provide resources for the software development process: testing, integration, validation, and benchmarking.
- Type of component: Service
- Owner: PSNC
- Expected TRL: 9
- Licensing: Proprietary
This component will allow for the effective use of the HPC infrastructure by: 1.providing the means for accessing HPC in terms of authentication, resource discovery, and job 2.managing an efficient way to run hybrid, cloud-based applications within the HPC environment accessing GPU and specialised low-latency networks and high throughput storage. 3.provisioning and orchestration of resources among multiple infrastructures.
- Type of component: Service
- Owner: PSNC
- Expected TRL: 9
- Licensing: Proprietary
The tool’s key advantages are its highly rated user-friendliness and the flexibility it provides to even a non-IT user to customise its templates according to their needs in an efficient and intuitive way. Its availability, reliability, and operational performance are of course very important as well since they are the foundation of a competitive tool.
- Type of component: Software
- Owner: ZELUS
- Expected TRL: 6
- Licensing: Apache 2.0
MARVEL corpus will give the possibility for SMEs and start-ups to build on top of these data assets and create new business by exploring extreme-scale multimodal analytics. Furthermore, by adopting an SLA enabled Big Data analytics framework, it is expected to maximise the impact that the MARVEL corpus will have on the international scientific and research community.
- Type of component: Service
- Owner: STS
- Expected TRL: 5
- Licensing: Public
IPR Management
There are 29 assets owned by various owners. In addition, although it is still early, as some partners have not yet decided under which type of license, they want to market their results, there is a perceived mix of Open Source licenses, but also Proprietary licenses, so a 100% Open Source approach will not be possible. In the future, the asset owners will consider the possibility of releasing parts of the developed technologies under MARVEL as Open Source approach. The Open Source licenses provided up to now, as MIT or Apache 2.0 do not present incompatibilities among them.
Funding
This project has received funding from the European Union’s Horizon 2020 Research and Innovation program under grant agreement No 957337. The website reflects only the view of the author(s) and the Commission is not responsible for any use that may be made of the information it contains.
