Cold Case Anne Frank: Artificial Intelligence for Good

The AI for Good team wanted to help discover valuable insights to help in the investigation of the Anne Frank cold case.

We developed a knowledge mining solution to help in the production of investigation and news program based on Artificial Intelligence able to create a relationship between all indexed resources such as documents, images, videos, and audios.

Artificial Intelligence
Azure Search, Azure Machine Learning, Cognitive Services.
We partnered with Microsoft and the investigator team to create a knowledge mining solution to help production of investigation and news program.
We developed a web application and search engine powered by Azure Search, Cognitive Services such as Translation, Language detection, face detection, OCR, and Azure Machine Learning to extract semantic information from the text and Video Indexer.
- Azure Search to retrieve data resources related to search queries and create semantical relationships.
- Enriched data with facial recognition, video transcription, advanced OCR to recognize handwriting, and keyphrase extraction.


The challenge we faced was to index all the content they have, initially more than 300,000 documents, to be able to perform intelligent searches using their own cognitive services and semantic models to create a relationship between all the different resources available to investigate the case: documents, images, videos and audios.


Thanks to the development of this project we have managed to assemble a very complex architecture that uses all the services of Cognitive Search, in addition to making a specific implementation using an IR algorithm.


Finally, more than 500,000 documents were indexed and since the researcher team started to work with our solution in September 2019, they have found 5 new relationships between victims and suspects.