Tomato Quality Classification Based on Transfer Learning Feature Extraction and Machine Learning Algorithm Classifiers

This study presents a robust method for classifying tomato quality using transfer learning feature extraction and machine learning classifiers. The process begins with the collection of a diverse tomato dataset, followed by pre-processing steps including image resizing, noise removal, contrast enhancement, and segmentation techniques to isolate the tomato from the background. Augmentation techniques such as rotations and translations are applied to increase the dataset's variability. Features are then extracted using the pre-trained InceptionV3 Convolutional Neural Network (CNN), leveraging its powerful capabilities for image feature representation. These features are subsequently used to train a Support Vector Machine (SVM) classifier, categorizing tomatoes into four quality classes: Damaged, Old, Ripe, and Unripe. The model's performance is evaluated using metrics such as accuracy, confusion matrix, F1 score, precision, recall, and specificity, demonstrating the system's efficacy in distinguishing between different tomato qualities. The proposed method aims to provide an efficient and accurate tool for automated tomato quality assessment, which is crucial for both agricultural production and market distribution.

Keywords: Tomato Dataset, Features Extraction, Image Processing Techniques, Machine Learning Algorithm, Deep Learning algorithm, and Accuracy.