A CNN-based Model to Classify Fresh and Damaged Fruit

Since there are no practical ways to deal with the wastage of food at distinct levels, including families, restaurants, and food supply chains, the amount of meals in the world is rapidly decreasing. Cooking too much and additional elements are to blame for a large bulk of food wastage in both homes and restaurants. We distinguished between fresh and damaged food to decrease food waste, which is mostly produced by families. If the flaws are not found, a diseased apple may infect a healthy one. Food waste is therefore more likely to occur, which leads to several issues. To distinguish between healthy and damaged fruits, input photos are employed. This study included a variety of fruits, including apples, bananas, and oranges. CNN collects fruit picture attributes, while SoftMax classifies images into fresh and rotting fruits. An ensemble strategy that integrates various convolutional neural networks is suggested in this to increase the accuracy rate. we designed and implemented a new CNN model from scratch to classify fresh and rotten fruits, after studying existing architectures M1, M2, M3, M4, M5, and M6. The present work is ensembling existing models and building the best accuracy model out of all combinations. These models are tested on a single data set obtained from Kaggle.com, and the recall, F1 Score, and F1 Score values are recorded. The efficacy of the proposed model was assessed using a dataset from Kaggle, and it showed a 98.08 percent accuracy rate.