This document discusses different methods for classifying flowers and counting petals from images using machine learning. It first reviews prior work on flower classification using both non-deep and deep learning techniques. It then outlines the goals of attempting to count petals from images and classify flowers in a single step model. The methodology section describes using k-means segmentation, edge detection, and other image processing techniques to count petals, and training a ResNet50 and LSTM model for classification. Results show the deep learning model achieved 90% accuracy while the petal counting methods were less accurate. The conclusion discusses challenges with petal counting and opportunities for future work in deep instance segmentation and improving the flower datasets and images.