A Study on Flower Classification Using Deep Learning Techniques

Abstract

A vital component of the ecological order, the flower is one of the plant's organs. There are various applications for flowers that are advantageous to people. Flowers today come in more than 400000 different types. Due to the similarities in shape and color of different flower types, it is challenging to distinguish them from one another. The wide variety of shapes, the distribution of colors, the lighting, and the distortion of exposure make flower classification a difficult challenge. With some photos, it gets harder for the human eye to distinguish between flowers that are similar in color and shape. For picture recognition and classification issues, deep learning algorithms are frequently used. Deep learning architectures have been improved throughout time to include more layers and create more reliable models for categorization issues. Researchers recently used CNN models to solve a variety of classification issues, doing away with the necessity for manual elements of classification. In this study, Deep CNN-based traditional artificial neural networks are used for image classification and identification of flower species. Popular pre-trained learning techniques such as VGG19, RCNN, Fast R-CNN, GoogleNet, and ResNet are conducted to classify flower species.