Dataset for multiclassification of parts in industry

Data used with 5 different parts for iterative reference set creation. Images were taken with a Logitech camera from parts that were positioned on a table in different ways. It contains two folders, a development one and a test one, randomly built. This is a small dataset that suits the purpose of not using lots of data for multiclassification, which was the objective of the present work. It was shown that is possible to acquire just some relevant images from each in order to train a classifier, a feature descriptor in this case.

搜集汇总

背景与挑战

背景概述

该数据集是一个工业零件多分类的小型图像数据集，包含5种不同零件，通过Logitech相机拍摄，零件以不同方式放置。数据集分为开发集和测试集，随机构建，旨在支持多分类任务，并证明使用少量相关图像即可训练分类器（如特征描述符）。数据集适用于图像处理和分类研究，遵循CC BY 4.0许可。