NEU surface defect database ( classification )

该数据集是东北大学的宋克臣等几位老师收集的，NEU surface defect database数据集收集了夹杂、划痕、压入氧化皮、裂纹、麻点和斑块6种缺陷，每种缺陷300张，图像尺寸为200×200。我已经分成了270*6张训练集图片和30*6张测试集图片

This dataset, named the NEU Surface Defect Database, was collected by Professor Song Kechen and other faculty members from Northeastern University. It includes six types of surface defects: crazing, inclusion, patch, pitting surface, scratch, and rolled-in scale. Each defect type has 300 images with a size of 200×200 pixels. The dataset has been pre-split by me into a training set with 270 images per defect (totaling 270*6 images) and a test set with 30 images per defect (totaling 30*6 images).

搜集汇总

背景与挑战

背景概述

该数据集是东北大学收集的用于表面缺陷分类的数据库，包含夹杂、划痕、压入氧化皮、裂纹、麻点和斑块6种缺陷类型，每种缺陷有300张200×200像素的图像。数据集已预先划分为训练集和测试集，适用于机器学习分类任务。