ITML's Data Fusion Bus (DFB), is a customizable product for data fusion for multiple modality data streams. It is using a modality processing pipeline that can transform heterogeneous data in real time by reducing the granularity from the signal/raw level to a semantic level.
The DFB includes autonomous fusion agents that are activated during the execution of the modality processing pipeline to perform modality fusion techniques on the data under transformation.
Α user interface is currently under development, to allow configuration, execution and visualization of the fusion process.
The DFB is highly scalable as it is built on the Apache Kafka distributed streaming platform.
Unique advantages of DFB are: achieve semantic fusion of events not coinciding in time, very accurate synchronisation between image/faceID and network/entityID, distribution mechanism interchanging data between the modules running on different machines, and supporting common interfaces to share data to third party services.