Project information

• Category: Computer Vision - Deep Learning
• Done for: University Module
• Project date: 2023
• Project URL: https://github.com/Daniel2tio/Segment_Class_CNN

This paper explores utilizing pre-trained convolutional neural networks (CNNs) for flower classification and developing a custom CNN model for flower segmentation, highlighting their importance in botanical research, agriculture, and environmental monitoring amidst recent advancements in computer vision.

This project explores using deep learning, specifically convolutional neural networks (CNNs), for flower classification and segmentation. The study aims to benefit botanical research, agriculture, and environmental monitoring. The project utilizes transfer learning with the GoogleNet CNN model for flower classification and a custom CNN model for flower segmentation. It discusses the limitations of traditional methods and the advantages of CNNs, along with data augmentation techniques. Evaluation results show promising accuracy in both classification and segmentation tasks. Suggestions for improvement include increasing dataset diversity and fine-tuning model architectures. Overall, the project demonstrates the effectiveness of CNNs in flower analysis, offering potential applications in automated flower identification and analysis processes.