AI Robotphoto
Generating photos automatically for the teaching phase of image recognition AI algorithms
Factsheet
Client
Own R&D
Date
Summer 2019
Project
3D AI Merging
Technology
3D/ML
Platform
Unity
Car recognition
If the task is to prepare an app, which recognizes a certain model of a certain car brand on the image – this is the solution.
Sign Recognition
Also, if the object to recognise is a wet floor sign – our system is there to help, too.
Video
Insight
The most popular areas of AI is image recognition, speech recognition and big data analytics. In the core of all these areas you can find the essence of AI – the algorithms. These are the products of the AI systems – which need huge amount of data to come up with these algorithms. In the optimal cases these data is already available from some source – years of recordings, saving transaction logs, sensor data, etc. In the case of image recognition this is not that simple. To make the an object recogniser AI you need thousands of images with that object on it and the question is how will you prepare those – making them yourself with a camera in the hand and placing the object in numerous places. This seems to be unreal. Our POC projects offers a solution to this problem by preparing these photos automatically. It uses the 3D model of the object and places it at various orientations on regular images.
