- A 5 layer CNN model to detect plant disease using leaf images.
- The model is trained on Plant Village Dataset.
- The dataset has 61,486 images of total 39 categories of 16 variety plants and backgrounds.
- The dataset has been augmented using different techniques like image flipping, Gamma correction, noise injection, PCA color augmentation, rotation, and Scaling.
-
The model consisted of 9 Convolutional layers followed by ReLU, Batch Normalization and Max Pooling.
-
80% of dataset has been used for training, 10% for validation and 10% for testing.
-
The accuracy of the model on training set is 93.07%, validation set is 91.53% and test is 90.94%.
-
References: