GRAPH-BASED IMAGE RECOGNITION

ABSTRACT
In the last few years, image recognition rapidly improved its capabilities both on object and people recognition. Anyway its real potential can remain unexploited without a sufficiently powerful infrastructure and suitable use-cases that can support data mining and ah-hoc insights. This tutorial will present a graph based image recognition system to show how an awesome synergy can be created leading to a wide set of practical applications in security and social analysis. The discussion will go through the design of a suitable knowledge graph, the Graph DBMS (Neo4J) that will be the backbone of the system and its querying language Cypher, the details in people and object fingerprints management (in and out the DB) and the usage of the Knowledge Graph Platform HUME for data visualization and manipulation. Under these perspectives we will show on place examples highlighting the possible insights that could arise from people to people and objects to people co-occurrences graph analysis.
SPEAKERS

Dr. Marco Del Coco

Dr. Marco Del Coco is a researcher in the National Council of Research (CNR) working for more than 10 years as a data scientist. He joined the research world as a PhD student enjoying the challenge of everyday new technologies and unseen ideas. Some day after the end of his PhD he was enrolled as research fellow in the CNR, moving his research interests on image processing and ML. Anyway, feeling the lack of a company experience, he successively joined Engineering and then Graphaware growing his experience on result/product-oriented research. In the last period he went back in accademia where he is trying to bring his effort combining all his past experiences.

 

Dr. Alessandro Negro

Dr. Alessandro Negro is the Chief Scientist at GraphAware. He is the author of Graph-Powered Machine Learning (Manning, 2020) and a thought leader in the space of graph-based AI and ML. He is one of the main authors of Hume, GraphAware’s flagship product and a leading Knowledge Graph Platform. He specialises in natural language processing, recommendation engines, fraud detection, and graph-aided search. Before joining the team, Alessandro gained over 10 years of experience in software development and spoke at many prominent conferences such as JavaOne. Alessandro holds a Ph.D. in Computer Science from University of Salento. He is based in Southern Italy (lucky him!) but travels to clients around the world.