Description of Invention
This research introduces a novel approach to identifying corn leaf disease, focusing on sustainable agriculture in Malaysia. We propose an end-to-end transfer learning model that reduces parameters while improving disease identification accuracy. The method leverages pre-trained CNNs, specifically DenseNet121 and InceptionV3, to extract deep features from corn plant images, comparing their performance. Our approach considers parameter efficiency in distinguishing between healthy and diseased corn leaves. We use DenseNet121 and InceptionV3 to extract deep features, which are then combined through concatenation, creating a richer feature set that enhances learning and accuracy.