How would AI assist a criminal investigation?

This month, Juliette, a 14 year old student, did her end of middle school internship at Eden AI. As she wants to solve crimes in the future, she decided to link her aspiration with what she was able to see during her one-week internship. In order to do that, she has created a police investigation where she tried to find a solution with the help of AI technologies. Here is how she integrated Eden AI and its features in her crime solving mystery.

‍

The investigation follows a crime committed in London. At the crime scene, the police take pictures of the room in which the crime took place. In order to identify everything they can see in this picture, they use a technology called Object Detection. This technology will be used to get a result of every object present in the picture.

‍

In the picture of the objects present at the crime scene, a note catches the police attention. As it is written in Japanese, they can’t read it. This step will go through two processes: use Optical Character Recognition (OCR) to extract the text from the letter and put it on the computer. Then, use Automatic Translation to translate for us what the letter says. The copy of the original text will be pasted into the area provided for this purpose; the result is the translation of the original text.

‍

The letter is long and the police are running out of time. As a result, they use Keywords Extraction in order to primarily focus on the important information that could be relevant to the investigation.

‍

While looking through the victim’s phone, they come across 118 pictures and videos of landscapes that they analyze thanks to Artificial Intelligence. They decide to keep the videos taken at night and discover 4 videos taken at a port, a few hours before the crime was committed. In these videos, we can observe four men dressed in leather jackets with a logo strapped on them. In order to try to find what this logo represents, the police use a Logo Detection technology that allows them to find out that this is a gang-affiliated symbol.

‍

Their investigation leads them to the port and they make the correlation between the containers present there. They take pictures of all the containers to try to find something that can help them. OCR is used again this time to see if they find the same logo that was on the jackets, on the containers.

‍

As they continue their investigation, they find a mountain's drawing. They use Landmark Detection to analyze the drawing and find a mountain that looks like this one in the world.

‍

‍

The investigation ends after the arrest of the gang and mainly of the two main killers who end up being identified thanks to Face Detection.

‍

Juliette showed us an investigation made possible by Artificial Intelligence technologies. These technologies are part of different domains (vision, text, translation, etc.) which requires the use of different APIs to run them. Here, Juliette wanted to draw a parallel with Eden AI, by showing the need to have an interface capable of offering all these technologies, instead of using different single interfaces for each of them. Eden AI acts as an aggregator of AI solutions under a single platform. With Eden AI, it is possible to use different technologies simultaneously and to combine them with each other as we house several AI technologies on our platform. In this scenario, the police could have contacted us to gather on a single API, all the technologies they needed to solve their investigation.

Stay tuned to see if Eden AI is ever going to be used in a police investigation case. In the meantime, you can contact us if you have any questions you need to ask about our features or our API!