Abstract:-

The immense deforestation happening in the recent years led to the extinction of many tree species. It is thus important to preserve details of the trees. Leaves are easy to collect and can tell many details regarding the trees. We can use image classification to identify the leaves. The development of deep convolution network has been a major breakthrough in image classification. However, it is very difficult to develop and train a custom CNN, this is where transfer learning comes into the picture. Transfer Learning uses a pre trained model which was used to solve one problem be reused to solve another problem.

In this study, we use pre trained models like ‘ResNet 50’ and ‘VGG 16’ along with classification algorithms like Decision Tree, Random Forest, Logistic Regression, KNN, Support Vector Machine (SVM) to classify images. Among all the models, SVM gave the best accuracy of 93% with VGG16 and Logistic Regression has given the best accuracy of 93.5% with ResNet 50.

Results: