Inicio  /  Computation  /  Vol: 10 Par: 8 (2022)  /  Artículo
ARTÍCULO
TITULO

A Study on Sample Size Sensitivity of Factory Manufacturing Dataset for CNN-Based Defective Product Classification

Dongbock Kim    
Sat Byul Seo    
Nam Hyun Yoo and Gisu Shin    

Resumen

In many small- and medium-sized enterprises (SMEs), defective products are still manually verified in the manufacturing process. Recently, image classification applying deep learning technology has been successful in classifying images of defective and intact products, although there are few cases of utilizing it in practice. SMEs have limited resources; therefore, it is crucial to make careful decisions when applying new methods. We investigated sample size sensitivity to determine the stable performance of deep learning models when applied to the real world. A simple sequential model was constructed, and the dataset was reconstructed into several sizes. For each case, we observed its statistical indicators, such as accuracy, recall, precision, and F1 score, on the same test dataset. Additionally, the loss, accuracy, and AUROC values for the validation dataset were investigated during training. As a result of the conducted research, we were able to confirm that, with 1000 data points or more, the accuracy exceeded 97%. However, more than 5000 cases were required to achieve stability in the model, which had little possibility of overfitting.

 Artículos similares

       
 
Daniel Suarez-Riera, Luca Lavagna, Juan Felipe Carvajal, Jean-Marc Tulliani, Devid Falliano and Luciana Restuccia    
Biochar, the solid sub-product of biomass pyrolysis, is widely considered an effective water retention material thanks to its porous microstructure and high specific surface area. This study investigates the possibility of improving both mechanical and r... ver más
Revista: Applied Sciences

 
Huizhong Ma, Liandi Wang, Na Li, Junpu Li and Lan Zhang    
Diamond-like carbon (DLC) has attracted much attention due to its unique properties such as high chemical inertness, optical transparency, and high biocompatibility. In this study, the total gas flow rate was kept constant, while the ratio of reactive ga... ver más
Revista: Applied Sciences

 
Mengping Huang, Shuai Ma, Jinrong He, Wei Xue, Xueyan Hou, Yuqi Zhang, Xiaofeng Liu, Heping Bai and Ran Li    
Amino acids found in minor coarse cereals are essential for human growth and development and play a crucial role in efficient and rapid quantitative detection. Surface-enhanced Raman spectroscopy (SERS) enables nondestructive, efficient, and rapid sample... ver más
Revista: Applied Sciences

 
Ana Pérez-Escoda, Andrés Barrios-Rubio, Luis Miguel Pedrero-Esteban and Carolina Ávalos    
Today?s public sphere is increasingly shaped by a dynamic, global, cross-cutting digital landscape, mostly ruled by social media and algorithms. Individuals are the raw material, the product, in this digital scenario, insofar as they generate and create ... ver más
Revista: Information

 
Yusuf Brima, Ulf Krumnack, Simone Pika and Gunther Heidemann    
Self-supervised learning (SSL) has emerged as a promising paradigm for learning flexible speech representations from unlabeled data. By designing pretext tasks that exploit statistical regularities, SSL models can capture useful representations that are ... ver más
Revista: Information