Classification of Biscuit Quality With Deep Learning Algorithms

IF 3.4 2区农林科学 Q2 FOOD SCIENCE & TECHNOLOGY

Journal of Food Science Pub Date : 2025-07-17 DOI:10.1111/1750-3841.70379

Oya Kilci, Yusuf Eryesil, Murat Koklu

{"title":"Classification of Biscuit Quality With Deep Learning Algorithms","authors":"Oya Kilci, Yusuf Eryesil, Murat Koklu","doi":"10.1111/1750-3841.70379","DOIUrl":null,"url":null,"abstract":"<div>\n \n <section>\n \n <h3> ABSTRACT</h3>\n \n <p>This study aims to reduce time, costs, and human errors in quality control processes for biscuit production by employing deep learning models to detect defective products. Two datasets were created. One for binary classification (defect and no defect) and another for multi-class classification (overcooked, texture defect, and not complete). Among the tested models, EfficientNet achieved the highest performance, with 93.89% accuracy, 96.74% precision, and a 95.38% F1 score in binary classification, and 95.03% accuracy in multi-class classification. ResNet showed comparable performance with accuracy rates of 93.38% and 95.23% for the respective datasets. While XceptionNet and MobileNet exhibited slightly lower accuracy rates, they delivered competitive F1 scores, particularly in detecting texture defects. Grad-CAM visualizations highlighted EfficientNet's superior focus on critical defect regions, reinforcing its suitability for industrial applications. These findings demonstrate the potential of deep learning models for efficient and precise quality control in industrial food production.</p>\n </section>\n </div>","PeriodicalId":193,"journal":{"name":"Journal of Food Science","volume":"90 7","pages":""},"PeriodicalIF":3.4000,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Food Science","FirstCategoryId":"97","ListUrlMain":"https://ift.onlinelibrary.wiley.com/doi/10.1111/1750-3841.70379","RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"FOOD SCIENCE & TECHNOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

ABSTRACT

This study aims to reduce time, costs, and human errors in quality control processes for biscuit production by employing deep learning models to detect defective products. Two datasets were created. One for binary classification (defect and no defect) and another for multi-class classification (overcooked, texture defect, and not complete). Among the tested models, EfficientNet achieved the highest performance, with 93.89% accuracy, 96.74% precision, and a 95.38% F1 score in binary classification, and 95.03% accuracy in multi-class classification. ResNet showed comparable performance with accuracy rates of 93.38% and 95.23% for the respective datasets. While XceptionNet and MobileNet exhibited slightly lower accuracy rates, they delivered competitive F1 scores, particularly in detecting texture defects. Grad-CAM visualizations highlighted EfficientNet's superior focus on critical defect regions, reinforcing its suitability for industrial applications. These findings demonstrate the potential of deep learning models for efficient and precise quality control in industrial food production.

Abstract Image

查看原文本刊更多论文

基于深度学习算法的饼干质量分类

本研究旨在通过使用深度学习模型来检测缺陷产品，减少饼干生产质量控制过程中的时间、成本和人为错误。创建了两个数据集。一个用于二元分类（有缺陷和无缺陷），另一个用于多级分类（过熟、质地缺陷和不完整）。在测试的模型中，effentnet的准确率为93.89%，精密度为96.74%，在二元分类中F1得分为95.38%，在多类分类中准确率为95.03%。ResNet在各自数据集上的准确率分别为93.38%和95.23%。虽然XceptionNet和MobileNet的准确率略低，但它们提供了具有竞争力的F1分数，特别是在检测纹理缺陷方面。Grad-CAM可视化突出了EfficientNet对关键缺陷区域的卓越关注，加强了其对工业应用的适用性。这些发现证明了深度学习模型在工业食品生产中有效和精确的质量控制的潜力。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Food Science 工程技术-食品科技

CiteScore

7.10

自引率

2.60%

发文量

412

审稿时长

3.1 months

期刊介绍： The goal of the Journal of Food Science is to offer scientists, researchers, and other food professionals the opportunity to share knowledge of scientific advancements in the myriad disciplines affecting their work, through a respected peer-reviewed publication. The Journal of Food Science serves as an international forum for vital research and developments in food science. The range of topics covered in the journal include: -Concise Reviews and Hypotheses in Food Science -New Horizons in Food Research -Integrated Food Science -Food Chemistry -Food Engineering, Materials Science, and Nanotechnology -Food Microbiology and Safety -Sensory and Consumer Sciences -Health, Nutrition, and Food -Toxicology and Chemical Food Safety The Journal of Food Science publishes peer-reviewed articles that cover all aspects of food science, including safety and nutrition. Reviews should be 15 to 50 typewritten pages (including tables, figures, and references), should provide in-depth coverage of a narrowly defined topic, and should embody careful evaluation (weaknesses, strengths, explanation of discrepancies in results among similar studies) of all pertinent studies, so that insightful interpretations and conclusions can be presented. Hypothesis papers are especially appropriate in pioneering areas of research or important areas that are afflicted by scientific controversy.