Enhancing Plant Species Recognition: A Multi-attribute Deep Learning Approach

With the large number of plant species in the globe and the shortcomings of conventional techniques that rely on leaf features, this study tackles the problem of plant species identification. To improve plant recognition, it suggests a multi-attribute method that combines machine learning and botanical taxonomy. The project focuses on applying transfer learning to categorize plants using attributes other than leaves by using deep learning models like Inception ResNetV2, MobileNetV2, InceptionV3, and VGG16. After assembling a dataset of 12,000 photos covering three different plant species, the effectiveness and accuracy of the models were evaluated. Although the models were 100% accurate in leaf-based identification, they were not as good at other plant features.