TÊN TIẾNG ANH: BUILDING MODELS AND AN APPLICATION TO IDENTIFY FLOWERS

Abstract

In recent years, deep learning methods have become crucial for complex tasks, including feature extraction, segmentation, and classification of images. This impact is particularly notable in flower species classification. This research introduces a neural network model designed for image classification, utilizing a fully connected layer and softmax layer, along with the feature extraction part of the Inception-V3 model combined with Transfer Learning approach, the model aiming to classify species using on the DIY 30 genus of flower image dataset. The dataset was created by collecting images based on images of 30 flower genera in available datasets. We employ techniques for image augmentation and image processing to enhance the efficiency of extracting flower images. Then, I created an application that can deploy the model so that it can be easily used by users to classify the genus of flower they desire. The mean testing accuracy of 76.8% highlights the adaptability and efficacy of deep learning architectures across diverse applications, particularly in image classification. This outcome also lays the foundation for ongoing advancements in the field.