LEGO bricks are highly popular due to the ability to build almost any type of creation. This is possible thanks to availability of multiple shapes and colors of the bricks. For the smooth build process the bricks need to properly sorted and arranged. In our work we aim at creating an automated LEGO bricks sorter. With over 3700 different LEGO parts bricks classification has to be done with deep neural networks. The question arises which model of the available should we use? In this paper we try to answer this question. The paper presents a comparison of 28 models used for image classification trained to classify objects to high number of classes with potentially high level of similarity. For that purpose a dataset consisting of 447 classes was prepared. The paper presents brief description of analyzed models, the training and comparison process and discusses the results obtained. Finally the paper proposes an answer what network architecture should be used for the problem of LEGO bricks classification and other similar problems.
Authors
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1007/978-3-031-08757-8_52
- Category
- Aktywność konferencyjna
- Type
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Language
- angielski
- Publication year
- 2022
