Flower Classification Using Deep Learning by using an Web-based API for Accurate and User-Friendly Identification

Abstract

It is important to classify flowers because it forms a major part of botanical studies, agriculture and study of the environment among others. The idea to be developed in this project can be referred to as Flower Classification with Deep Learning where rather than describing the species in detail, its identification on the basis of image data with the help of deep learning algorithms will be the main goal. Through convolutional neural networks (CNNs) of this project, a user can classify flowers by entering an image into the API interface created using Streamlit. The work started by capturing the flower image dataset, data enhancement including, changing contact, resizing of the input images of size 224×224, adjusting the gamma values of the images and data augmentation. The findings confirm the promises of deep learning in tackling challenging visual classifiers and offer a growing point for improvement for future work on self-driving plant species identification. This work is able to pragmatically serve as a tool for classifying flowers in a way that makes this activity considerably less time-consuming than it has been hitherto; therefore, this work is recommended to all those who focus on plant sciences, whether they are researchers, teachers, or amateurs.