Project information

• Category: Machine learning
• Client: University of Kelaniya
• Project URL: project link

This project was focused on exploring the use of Tranfer learning to classify diseases in tea leaves.

Transfer learning is highly advantageous because it capitalizes on the knowledge acquired from one task and applies it to another, often related, task. This approach leverages pre-trained models that have been trained on vast datasets, enabling them to learn intricate patterns and representations in data. By fine-tuning these models on a new, target task with a relatively smaller dataset, transfer learning accelerates the training process and improves overall performance. This is particularly beneficial in scenarios where collecting a large dataset for a specific task might be impractical due to time, cost, or availability constraints. Transfer learning also aids in handling the problem of overfitting, as the model has already learned general features from a diverse range of data. This fosters the creation of models that are more adaptable and capable of better generalization, even in domains where labeled data is scarce.

The use of Transfer learning has shown to result in remarkable performances in image classification. The project uses the popular Resnet50 model.