A Transfer Learning Based Approach To Detect Rotten Dragon Fruits

Abstract

Transfer learning is a promising approach for improving the accuracy of detection of rotten dragon fruits. In this paper, a transfer learning-based approach is proposed for detecting rotten dragon fruits using images of dragon fruit from different places in Bangladesh. The proposed approach uses a pre-trained network which is fine-tuned on a dataset containing images of rotten and fresh dragon fruits. The dataset is created by collecting images of dragon fruits from different places in Bangladesh. The pre-trained network is fine-tuned on the dataset, and the accuracy of the detection is improved by using transfer learning. The results of the experiments show that the proposed approach achieves an accuracy of up to 98.57% for detecting rotten dragon fruits. Furthermore, the proposed approach is highly robust and can detect rotten dragon fruits irrespective of their color, shape, texture and other features. The results of the experiments demonstrate that the proposed approach is effective and can be used for detecting rotten dragon fruits in a cost-effective manner.