Classification of Rotten Fruits vs Fresh Fruits Using Sequential Model with Convolutional Neural Network

This research employs Convolutional Neural Networks (CNNs) to distinguish between ripe and spoiled fruits using deep learning. The CNN model achieves an impressive accuracy rate of 98.79^% in accurately categorizing these two groups. To facilitate model training and evaluation, a comprehensive dataset containing images of both fresh and rotten fruits is utilized. Libraries such as TensorFlow, OpenCV, and matplotlib are employed for efficient dataset processing. The CNN architecture is thoughtfully crafted, incorporating dropout layers to address overfitting and capture intricate fruit image characteristics. A novel approach for categorizing fruit spoilage is introduced, and the study focuses on apples, bananas, and oranges. The model's performance is evaluated using a Kaggle-sourced dataset, yielding an accuracy rate of 98.0% (0.9879), highlighting the effectiveness of the CNN model in discerning between ripe and spoiled fruits. Additionally, the research explores transfer learning techniques for fruit classification.