Cassava Leaf Disease Classification using Deep Neural Networks

Abstract

In recent years, deep learning has gained much popularity over traditional machine learning techniques in terms of accuracy and precision when trained on substantial amount of data. In this work, a state-of-the-art deep learning technique has been employed for classification and prediction of cassava leaf diseases. Being the second largest producer of carbohydrates in the world, cassava plant has become an important source of calories for people in tropical regions, but it is highly susceptible to viral, bacterial, and fungal attacks resulting in stunted plant growth and hence the yield. So, the aim of the research is to help the farmers quickly identify diseased leaves before they cause any severe damage. The dataset that is used in this work is taken from Kaggle competition 2020 containing 21,397 images of cassava plant leaves belonging to 5 classes: Cassava Bacterial Blight, Cassava Brown Streak Disease, Cassava Green Mottle, Cassava Mosaic Disease and Healthy leaves. In this work, EfficientNet model B4 was trained using transfer learning approach. Further, to remove background noise, Segmentation was performed using U-Net to extract only the leaves from images. Our system provided reasonable performance when validation data was provided to trained model yielding 81.43% and 89.09% accuracy on original and segmented datasets, respectively.