This document presents a methodology for classifying flower images using transfer learning. It discusses how transfer learning can help overcome issues with lack of labeled data by leveraging knowledge gained from solving similar problems. The methodology involves pre-processing images, extracting features, fine-tuning a pre-trained model with the flower data, and evaluating results with metrics like confusion matrices and ROC curves. A literature review covers previous work on flower classification using techniques like color, texture, histograms, neural networks and convolutional neural networks, with transfer learning achieving better results than conventional approaches. The objectives are to classify different flower species using transfer learning and analyze model performance.