Flower perception using Convolution Neural Networks based Escalation of Transfer learning

Flowers are one of the most beautiful creations ever made by god, with a plenty of different varieties and hues to choose from. A botanist with extensive knowledge and abilities is required to identify each of them. As there is a very large number of species of flower which resemble the same in shape, look, or surrounding items like the shape of the leaves and the grass, flower categorization is a difficult undertaking? Flower recognition is a significant problem that can be solved manually by specialists, but it takes a long time and is consequently inefficient. As a result, automatic plant recognition is a vital topic of study. In this technologically advanced era, artificial intelligence is being used to solve real-world problems, making the seemingly impossible conceivable. In this study, using CNN with transfer learning to improve floral recognition ability is suggested. The dataset contains 3670 photos of various flowers, with 2934 (75%) images used for training and 736 (25%) images used for testing. The images are divided into 26 classes. Experiments with a dataset of flowers reveal that the suggested model outperforms alternative classification approaches such as CNN, Alex Net, and Google Net.