Abstract

This report explores a methodology for flower classification using a pre-trained ResNet model. The approach involves fine-tuning the ResNet model on a flower dataset to achieve high classification accuracy. Key processes include data preprocessing, augmentation, model training, and evaluation, resulting in a validation accuracy of 95%. The project "Flower Classification Using Pre-trained ResNet Models in Computer Vision" aims to explore the capabilities of advanced deep learning techniques in accurately classifying flower species. Leveraging pre-trained Residual Networks (ResNet), this study demonstrates the efficacy of transfer learning in the domain of image classification, particularly for flora identification. This abstract outlines the methodology, key findings, and potential applications of the project, emphasizing the significant improvements in classification accuracy and computational efficiency achieved through the use of pre-trained models.