Scholarlink

Guava (Psidium guajava L.) is a tropical fruit native to Central America, known for its rich vitamin C content, which functions as an antioxidant and aids in the treatment of ailments such as mouth ulcers, swollen gums, and bleeding. Currently, the classification of guava quality is commonly performed through manual inspection of the fruit's exterior. This conventional method of classification often yields inaccurate and inconsistent results due to human error. Inaccurate grading can adversely affect farmers, as all fruit may be priced uniformly regardless of quality. The present study offers a classification model using Deep Learning techniques for grading the quality of guava fruit. We used Convolutional Neural Networks (CNNs) in this study for image classification. The dataset is comprised of 1312 images, with 649 for high-quality guava and 663 for low quality guava. This model was trained for 25 epochs getting a training accuracy of 0.96. The model attained an average accuracy of 0.96 while testing, with a precision score and recall of 0.98 and 1.00 respectively. The results show that the proposed model is very efficient in determining guava fruit quality, attaining accurate classifications of high- and low-quality samples.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Learning for Classifying Guava Fruit Quality Based on Guava Fruit Images

  • 章节

  • 发表日期：

    2026年

查看全文
  • Hasanatul Fu’adah Amran,
  • Andes Fuady Dharma,
  • Harun Mukhtar,
  • Rizka Hafsari,
  • Syahril,
  • Bayu Anugerah Putra,
  • Rahmad Fadila,
  • Rahmad Firdaus

摘要

Guava (Psidium guajava L.) is a tropical fruit native to Central America, known for its rich vitamin C content, which functions as an antioxidant and aids in the treatment of ailments such as mouth ulcers, swollen gums, and bleeding. Currently, the classification of guava quality is commonly performed through manual inspection of the fruit's exterior. This conventional method of classification often yields inaccurate and inconsistent results due to human error. Inaccurate grading can adversely affect farmers, as all fruit may be priced uniformly regardless of quality. The present study offers a classification model using Deep Learning techniques for grading the quality of guava fruit. We used Convolutional Neural Networks (CNNs) in this study for image classification. The dataset is comprised of 1312 images, with 649 for high-quality guava and 663 for low quality guava. This model was trained for 25 epochs getting a training accuracy of 0.96. The model attained an average accuracy of 0.96 while testing, with a precision score and recall of 0.98 and 1.00 respectively. The results show that the proposed model is very efficient in determining guava fruit quality, attaining accurate classifications of high- and low-quality samples.