Flower recognizer for mobile application

The aim of the project was to create a model suitable in a mobile context that can recognize flowers from images. To achieve this goal a flower dataset consisting of 102 classes was used. Three different fine-tuned architectures were compared and then they were compressed using iterative pruning and quantization. Analysing the results obtained, the final considerations were made, choosing the best model taking into account two fundamental aspects for the selected context: accuracy and size of the model.

Requirements

To install the requirements:

pip install -r requirements.txt

Data

The dataset used, called Oxford Flower Dataset, contains images of common flowers in the UK, belonging to 102 different categories and it can be downloaded from the dedicated website https://www.robots.ox.ac.uk/~vgg/data/flowers/102/index.html

References

• M.-E. Nilsback, “An automatic visual Flora – segmentation and classification of flowers images,” Ph.D. dissertation, University of Oxford, 2009. [Online]. Available: https://www.robots.ox.ac.uk/∼vgg/publications/2009/Nilsback09/
• M. Sandler, A. G. Howard, M. Zhu, A. Zhmoginov, and L. Chen, “Inverted residuals and linear bottlenecks: Mobile networks for classification, detection and segmentation,” CoRR, vol. abs/1801.04381, 2018. [Online]. Available: http://arxiv.org/abs/1801.04381
• A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam, “Mobilenets: Efficient convolutional neural networks for mobile vision applications,” CoRR, vol. abs/1704.04861, 2017. [Online]. Available: http://arxiv.org/abs/1704.04861
• G. Huang, Z. Liu, and K. Q. Weinberger, “Densely connected convolutional networks,” CoRR, vol. abs/1608.06993, 2016. [Online]. Available: http: //arxiv.org/abs/1608.06993
• M. Tan and Q. V. Le, “Efficientnet: Rethinking model scaling for convolutional neural networks,” CoRR, vol. abs/1905.11946, 2019. [Online]. Available: http://arxiv.org/abs/1905.11946
• I. Goodfellow, Y. Bengio, A. Courville, and Y. Bengio, Deep learning. MIT press Cambridge, 2016, vol. 1, no. 2, ch. 7.12, p. 253.

Authors

• Lorenzo Pirola      

• Matteo Romanato      

• Youssef Karrati