Glass Bottle Defect Dataset for Real-Time Industrial Inspection

This dataset contains industrial glass bottle images for real-time visual defect inspection. The dataset includes 1,479 images categorized into three classes: non-defective, large defect, and small defect. It was prepared to support research on automated glass bottle defect classification under practical industrial inspection conditions. The dataset was used in the study “M3XGB: Multi-Model Modular Transfer Learning for Real-Time Glass Bottle Defect Inspection.” It supports evaluation of deep learning and transfer learning models for classifying glass bottle defects, including comparison of classification accuracy, MCC, and deployment-oriented performance. Dataset classes:- Non-defective: 1,098 images- Large defect: 180 images- Small defect: 201 images The dataset is released under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0).