Evaluation of deep learning architectures applied to identification of diseases in grape leaves

• Flávio R. S. Oliveira IFPE
• Felipe C. Farias IFPE
• Bernardo João de Barros Caldas Business Consulting

DOI: https://doi.org/10.5753/eniac.2018.4447

Resumo

Vale do São Francisco in Pernambuco is one of the most economically important poles in the state and among its cultivars, it is worth mentioning the grape culture. This sector faces challenges related to the response time between identifying a field infestation and taking corrective actions, in order to minimize losses. This work comprises a comparative analysis between deep learning architectures, applied to identification of diseases in grape cultivars. Results suggest that the use of these technologies is plausible to differentiate healthy grape leaves from leaves presenting one of three different types of diseases, obtaining near 100% accuracy in studied database using an architecture that can be employed in embedded devices.

Referências

Codevasf Ministério da Integração Nacional (2016) “Relatório de Gestão do Exercício de 2016”,
http://www2.codevasf.gov.br/empresa/auditoriainterna/processosdecontasanuais/relatorio_de_gestao_2016.pdf . Acessado em 20 de Junho de 2018.

Gschwend, D. (2016) “Zynqnet: An fpgaaccelerated embedded convolutional neural network”. Master’s thesis, Swiss Federal Institute of Technology Zurich (ETHZurich),2016.

Hughes, D. e Salathé, M. (2015) “An open access repository of images on plant health to enable the development of mobile disease diagnostics”, arXiv:1511.08060 .

Iandola, F. N. et al. (2016) “Squeezenet: Alexnetlevel accuracy with 50x fewer parameters and< 0.5 mb model size”, arXiv preprint arXiv:1602.07360, 2016. Kamilaris, A. e PrenafetaBoldú,

F. (2018) "Deep learning in agriculture: A survey", In: Computers and Electronics in Agriculture 147, p. 70–90

Krizhevsky A., Sutskever, I. e Hinton, G. (2012) “Imagenet classification with deep convolutional neural networks”, In: Advances in neural information processing systems, p. 10971105

Lee, S., Chan, C., Wilkin, P., Remagnino, P. (2015) “Deepplant: Plant identification with convolutional neural networks”, In: Proceedings of 2015 IEEE International Conference on Image Processing (ICIP)

Lins, E. A. e Rieder, R., (2017) “Uma metodologia de contagem e classificação de afídeos utilizando visão computacional”, In: Proceedings of 2017 SIBGRAPI Conference on Graphics, Patterns and Images.

Pan, S. J. et al. (2010) “A survey on transfer learning”, IEEE Transactions on knowledge and data engineering, v. 22, n. 10, p. 13451359, 2010.

Paszke, A., Gross, S., Chintala, S., Chanan, G., Yang, E., DeVito, Z., Lin, Z., Desmaison, A., Antiga, L. e Lerer, A. (2017) “Automatic differentiation in PyTorch”, In: NIPS 2017 Workshop.

Schmidhuber, J. (2015) “Deep learning in neural networks: An overview.”, Neural networks, v. 61, p. 85117, 2015.

Silva, D. J. e Faria, C. M. B. (1999) “Amostragem para análise foliar da videira”, Embrapa Semiárido, 1999 https://www.embrapa.br/buscadepublicacoes//publicacao/154374/amostragemparaanalisefoliardevideira. Acessado em 20 de Junho de 2018.

Sousa, A. L., Salame, M. F. A., Nascimento Filho, F. J. e Atroch, A. L. (2017) "Redes Neurais Convolucionais Aplicadas ao Processo de Classificação de Cultivares de Guaranazeiros". In: Proceedings of XIV Encontro Nacional de Inteligência Artificial e Computacional, p. 855864.

Szegedy, C. et al. (2016), “Rethinking the inception architecture for computer vision”, In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016. p. 28182826.

Too, E. C.,Yujian, L., Njuki, S. e Yingchun, L. (2018) "A comparative study of finetuning deep learning models for plant disease identification", In: Computers and Electronics in Agriculture.