Design and Implementation of a Flower Image Classification System Based on Convolutional Neural Network (CNN)

Hongru Li; Zhitao Wu

doi:10.23977/autml.2025.060208

Design and Implementation of a Flower Image Classification System Based on Convolutional Neural Network (CNN)

Download as PDF

DOI: 10.23977/autml.2025.060208 | Downloads: 0 | Views: 26

Author(s)

Hongru Li ¹, Zhitao Wu ¹

Affiliation(s)

¹ School of Electronic and Information Engineering, University of Science and Technology Liaoning, Anshan, China

Corresponding Author

Zhitao Wu

ABSTRACT

Flower image classification holds significant practical value in fields such as agricultural production, horticultural management, and ecological protection. Traditional flower classification methods based on manual feature extraction suffer from low efficiency and poor robustness, making them difficult to adapt to complex image scenarios. To address this issue, this paper designs and implements a flower image classification system based on Convolutional Neural Network (CNN). The system takes the public Oxford 102 Flowers dataset as the research object, and optimizes the quality of input data through preprocessing operations such as unified image size, normalization, and data augmentation. A lightweight CNN classification model is constructed with reference to the LeNet and simplified AlexNet structures, reducing the parameters of convolutional layers and fully connected layers to lower computational costs. Techniques including learning rate decay and batch normalization are adopted for model training and optimization. Comparative experiments are conducted to analyze the impacts of different optimizers (SGD vs. Adam) and data augmentation on the classification performance of the model. Experimental results show that the model with Adam optimizer and data augmentation achieves a classification accuracy of 92.3% on the test set, representing an increase of 15.6 percentage points compared with the unoptimized model. Characterized by simple structure, easy reproducibility, and excellent classification performance, this system provides a feasible reference for entry-level research on deep learning-based image classification.

KEYWORDS

Convolutional Neural Network; Image Classification; Flower Recognition; Data Augmentation; Model Optimization

CITE THIS PAPER

Hongru Li, Zhitao Wu, Design and Implementation of a Flower Image Classification System Based on Convolutional Neural Network (CNN). Automation and Machine Learning (2025) Vol. 6: 56-62. DOI: http://dx.doi.org/10.23977/autml.2025.060208.

REFERENCES

[1] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks[J]. Advances in neural information processing systems, 2012, 25.
[2] LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 2002, 86(11): 2278-2324.
[3] Kingma D P. Adam: A method for stochastic optimization[J]. arXiv preprint arXiv:1412.6980, 2014.
[4] Goodfellow I, Bengio Y, Courville A, et al. Deep learning[M]. Cambridge: MIT press, 2016.
[5] Ioffe S, Szegedy C. Batch normalization: Accelerating deep network training by reducing internal covariate shift[C]//International conference on machine learning. pmlr, 2015: 448-456.
[6] Srivastava N, Hinton G, Krizhevsky A, et al. Dropout: a simple way to prevent neural networks from overfitting[J]. The journal of machine learning research, 2014, 15(1): 1929-1958.

Subscription

E-Mail Alert

Downloads:	4329
Visits:	201672

Design and Implementation of a Flower Image Classification System Based on Convolutional Neural Network (CNN)

Author(s)

Affiliation(s)

Corresponding Author

ABSTRACT

KEYWORDS

CITE THIS PAPER

REFERENCES

RESOURCES

JOIN US

PUBLICATION SERVICES

CONTACT US