The proposed project aims to develop an algorithm of machine learning applied for components recognition. The machine will be designed for the automatic identification and classification of any component within specific groups (resistors, transistors, condensers, etc.)
The project initial phase will focus on the problem analysis to define the most suitable solution. The structure creation and the coding for the training and testing activities will follow. They consist of 1) analysis of the database with the automated routine so as to permit the self-learning of components distinguishing features and 2) experimentation of self-learning effectiveness with new data, evaluating the results, and fine-tuning the solution, with a particular focus on control parameters.
The project will be closed with the implementation and optimization of the algorithm for the ordinary use by the end-user.
Thanks to this technology, STC will obtain two main advantages for his cost engineering service: reduction of classification process time of electronic components (which is usually performed manually) and greater identification precision.