Project Overview
Transfer-learning study on apple, banana, guava and mixed Indian-fruit disease datasets, processed separately and together. Models were compared with accuracy and loss curves, confusion matrices, precision, recall and F1. In the reported runs VGG16 reached 98.5%, VGG19 99.0% and Xception 99.9% accuracy.
Business Problem
Early, accurate detection of fruit disease helps farmers and exporters act before a crop is lost.
