|
|
Registro Completo |
Biblioteca(s): |
Embrapa Uva e Vinho. |
Data corrente: |
16/12/2020 |
Data da última atualização: |
20/12/2022 |
Tipo da produção científica: |
Capítulo em Livro Técnico-Científico |
Autoria: |
MELO, R. F. de; LIMA, G. L. de; CORRÊA, G. R.; ZATT, B.; AGUIAR, M. S. de; NACHTIGALL, G. R.; ARAÚJO, R. M. |
Afiliação: |
RAMÁSIO FERREIRA DE MELO; GUSTAVO LAMEIRÃO DE LIMA; GUILHERME RIBEIRO CORRÊA; BRUNO ZATT; MARILTON SANCHOTENE DE AGUIAR; GILMAR RIBEIRO NACHTIGALL, CNPUV; RICARDO MATSUMURA ARAÚJO. |
Título: |
Diagnosis of apple fruit diseases in the wild with Mask R-CNN. |
Ano de publicação: |
2020 |
Fonte/Imprenta: |
In: CERRI, R.; PRATI, R. C. (Eds). Intelligent Systems. BRACIS 2020. Lecture Notes in Computer Science, vol 12319. Springer, 2020. p. 256?270. |
DOI: |
https://doi.org/10.1007/978-3-030-61377-8_18 |
Idioma: |
Português |
Conteúdo: |
A major challenge in image classification tasks using Machine Learning, and in particular when using deep neural networks, is domain shifting in deployment. This happens when images during usage are capture in different conditions from those used during training. In this paper, we show that despite previous works on the diagnosis of apple tree diseases using standard Convolutional Neural Networks displaying high accuracy, they do so only when no domain shift is present. When the trained model is asked to classify photos of apples taken in the wild, a 22% reduction in F1 score is observed. We propose to treat the task as a segmentation problem and test two different approaches, showing that using Mask R-CNN allows not only to improve performance in the original domain by 3%, but also significantly reduce losses in the new domain (only 6% reduction in F1 score). We establish segmentation as an important alternative towards improving diagnosis of apple tree diseases from photos. |
Palavras-Chave: |
Apple fruits; Deep learning; Instance segmentation. |
Categoria do assunto: |
F Plantas e Produtos de Origem Vegetal |
Marc: |
LEADER 01790naa a2200241 a 4500 001 2128251 005 2022-12-20 008 2020 bl uuuu u00u1 u #d 024 7 $ahttps://doi.org/10.1007/978-3-030-61377-8_18$2DOI 100 1 $aMELO, R. F. de 245 $aDiagnosis of apple fruit diseases in the wild with Mask R-CNN.$h[electronic resource] 260 $c2020 520 $aA major challenge in image classification tasks using Machine Learning, and in particular when using deep neural networks, is domain shifting in deployment. This happens when images during usage are capture in different conditions from those used during training. In this paper, we show that despite previous works on the diagnosis of apple tree diseases using standard Convolutional Neural Networks displaying high accuracy, they do so only when no domain shift is present. When the trained model is asked to classify photos of apples taken in the wild, a 22% reduction in F1 score is observed. We propose to treat the task as a segmentation problem and test two different approaches, showing that using Mask R-CNN allows not only to improve performance in the original domain by 3%, but also significantly reduce losses in the new domain (only 6% reduction in F1 score). We establish segmentation as an important alternative towards improving diagnosis of apple tree diseases from photos. 653 $aApple fruits 653 $aDeep learning 653 $aInstance segmentation 700 1 $aLIMA, G. L. de 700 1 $aCORRÊA, G. R. 700 1 $aZATT, B. 700 1 $aAGUIAR, M. S. de 700 1 $aNACHTIGALL, G. R. 700 1 $aARAÚJO, R. M. 773 $tIn: CERRI, R.; PRATI, R. C. (Eds). Intelligent Systems. BRACIS 2020. Lecture Notes in Computer Science, vol 12319. Springer, 2020. p. 256?270.
Download
Esconder MarcMostrar Marc Completo |
Registro original: |
Embrapa Uva e Vinho (CNPUV) |
|
Biblioteca |
ID |
Origem |
Tipo/Formato |
Classificação |
Cutter |
Registro |
Volume |
Status |
URL |
Voltar
|
|
Registros recuperados : 1 | |
1. | | SANTOS, M. E. R.; SOUZA, B. M. de L.; ROCHA, G. de O.; FREITAS, C. A. S.; SILVEIRA, M. C. T. da; SOUSA, D. O. C. de. Estrutura do dossel e características de perfilhos em pastos de capim-piatã manejados com doses de nitrogênio e períodos de diferimento variáveis. Ciência Animal Brasileira, v. 18, e-37547, p. 1-13, jul. 2017.Tipo: Artigo em Periódico Indexado | Circulação/Nível: B - 2 |
Biblioteca(s): Embrapa Pecuária Sul. |
| |
Registros recuperados : 1 | |
|
Nenhum registro encontrado para a expressão de busca informada. |
|
|