- Duration: 6 months
- Year: 2020
- Participants: JMA + MGP
- Digitalization and connectivity
Project objectives
In this case, the project will focus on the identification of key types for duplicating machines. The project will address the development of an AI-based key identifier/classifier that JMA will use for the key copier product it is developing.
MGEP will participate in the development of the key identification algorithm, while the development of the final application (i.e.: tablet, etc.) related to the key copying product will be the responsibility of JMA. This final product will make use of the developed identification algorithm, and MGEP will advise JMA on how to integrate it into its product.
Tasks performed
Phase 1: Preliminaries:
- Optimization of image capture conditions.
- Generation of training image set.
- Labeling of Training Set Images.
Phase 2: Key Type Identifier
- Training of profile reconstruction model based on Deep Learning.
- Ranking of similar keys: Taking as a reference the key profile obtained in T2.1, a ranking of the most similar keys will be made in the database of key references (more than 4000 types of keys).
