This is a Research Project worked on building the Swedish Leaf Classifier by using Transfer_Learning, ML Algorithms and DeepConv-Nets
Plant classification is one of the most foremost tasks for scientists, field guides, and others because plants have a significant role to play in the natural circle of life. Our problem statement revolves around three objectives:
- Showing the usage of Transfer Learning(TL) in classification models.
- Comparing the working of that model with different Machine Learning(ML) Algorithms
- Designing a dedicated CNN model for this leaf classification problem.
The dataset used for this experiment is the Swedish Leaf Dataset,available at https://www.cvl.isy.liu.se/en/research/datasets/swedish-leaf, which is a database of 15 different plant species with a total of 1125 leaf images.
Experimental results showed that Random Forest (RF) achieved a classification accuracy of 98.83% against other ML algorithms with a combination of Grayscale images, HSV color moments, hu moments, and haralick features. The ResNet50 model gave us the best accuracy of 99.85% compared to other models. A CNN convolves learned features with input data and uses 2D convolutional layers. We have built our own CNN model from scratch and managed to reach an accuracy of 98.04%.