|
Enhancing the efficiency of a convolutional neural network model for rice variety classification using a customized prewitt operator |
|---|---|
| รหัสดีโอไอ | |
| Creator | Jatsada Singthongchai |
| Title | Enhancing the efficiency of a convolutional neural network model for rice variety classification using a customized prewitt operator |
| Contributor | Suttipong Klongdee, Nattavut Sriwiboon |
| Publisher | Mahasarakham University |
| Publication Year | 2569 |
| Journal Title | Journal of Science and Technology Mahasarakham University |
| Journal Vol. | 45 |
| Journal No. | 3 |
| Page no. | 337-347 |
| Keyword | Rice variety, deep learning, convolutional neural networks, EfficientNetB0 |
| URL Website | https://li01.tci-thaijo.org/index.php/scimsujournal |
| Website title | Journal of Science and Technology Mahasarakham University |
| ISSN | 1686-9664 (Print), 2586-9795(Online) |
| Abstract | Rice variety classification is a crucial process in the agricultural industry, directly impacting product quality and production efficiency. This article presents a method for enhancing rice variety classification efficiency using a Convolutional Neural Network (CNN) integrated with a customized Prewitt operator for precise edge detection from grain images. This technique improves the accuracy of distinguishing between rice varieties with similar characteristics. The experimental results demonstrate that the CNN model combined with the Prewitt operator achieves a classification accuracy of 98.5%. This performance surpasses previous research using other methods, such as Sobel and Canny edge detection, which reported accuracies ranging from 85% to 96%. This article discusses the development stages of the CNN model and the Prewitt operator and provides a comparison with related works, highlighting its effectiveness across diverse environments. |
