Automobile Parts Classification Using Deep Learning-Based Ensemble Model

Complicated, decentralised supply chains, like those in the automobile industry, have to deal with a wide range of products. A significant number of unlabelled components must be located and confirmed. This is often done manually, which introduces the possibility of human error. Automatic recognition of automobile parts is an extremely useful skill, and it may be put to use in a variety of contexts, including the examination of automobile parts for quality, automatic assembly of those parts, counting items of different product types during packing, among others (Waquar Azam in Automobile-parts-classification, 2022 [1]). A convolutional neural network-based deep neural network architecture is employed in this study to categorise auto components. The ensemble deep learning architecture underwent training and evaluation using an openly accessible dataset that was made available through Kaggle. It consists of 689 RGB images focusing on 14 distinct components. Initially, the dataset was partitioned into two categories: training and testing samples. The proposed model reached an accuracy of 85.39%. This article demonstrates a categorisation mechanism that is customisable and sensor free. The classification of automotive parts has a tremendous potential for the automotive sector and contributes to many applications of car production, model verification, and automobile inspection systems, among others.