Contextualizing images with heterogenous data from multiple sources
Combining images and relevant, useful information from various data sources to increase usefulness of images for insight and analysis.
Images that have metadata attached can be contextualized with information from data sources such as unstructured PDF documents. Data contextualization and access to heterogeneous data from various data sources is complicated by lack of common ways of representing data semantically. Very often relevant data resides in silos where data is difficult to access, not to mention challenges in integration of the data across multiple silos. Each silo can have different structures, and some silos may be unstructured, meaning that extraction of relevant data is challenging. As a result, Enterprise Knowledge Graphs and Ontology-based Data Access have emerged as data management paradigms that can enable integration and interoperability between different data sources.
This MSc thesis aims at developing a framework for contextualization of images with heterogenous data from multiple sources.
- Identify techniques for data contextualization and interoperability, such as the paradigms for Enterprise Knowledge Graphs and Ontology-based Data Access.
- Develop a data contextualization framework that will enrich metadata of tagged images with information from relevant structured and unstructured data sources.
- Implement the framework in a software prototype (Seekuence).
- Evaluate the framework through a use case pilot in a current SINTEF innovation project (LIACI).
Expected Results and Learning Outcome
- Framework for data contextualization of images with relevant structured and unstructured data sources.
- Software prototype.
- Evaluation of prototype with real business use cases.
Basic knowledge of semantic technologies will be considered an advantage.