| |
|
|
|
Registro Completo |
|
Biblioteca(s): |
Embrapa Agricultura Digital. |
|
Data corrente: |
14/09/2021 |
|
Data da última atualização: |
14/09/2021 |
|
Tipo da produção científica: |
Artigo em Periódico Indexado |
|
Autoria: |
GONÇALVES, J. P.; PINTO, F. A. C.; QUEIROZ, D. M.; VILLAR, F. M. M.; BARBEDO, J. G. A.; DEL PONTE, E. M. |
|
Afiliação: |
JULIANO P. GONÇALVES, UFV; FRANCISCO A. C. PINTO, UFV; DANIEL M. QUEIROZ, UFV; FLORA M. M. VILLAR, UFV; JAYME GARCIA ARNAL BARBEDO, CNPTIA; EMERSON M. DEL PONTE, UFV. |
|
Título: |
Deep learning architectures for semantic segmentation and automatic estimation of severity of foliar symptoms caused by diseases or pests. |
|
Ano de publicação: |
2021 |
|
Fonte/Imprenta: |
Biosystems Engineering, v. 210, p. 129-142, Oct. 2021. |
|
DOI: |
https://doi.org/10.1016/j.biosystemseng.2021.08.011 |
|
Idioma: |
Inglês |
|
Conteúdo: |
Colour-thresholding digital imaging methods are generally accurate for measuring the percentage of foliar area affected by disease or pests (severity), but they perform poorly when scene illumination and background are not uniform. In this study, six convolutional neural network (CNN) architectures were trained for semantic segmentation in images of individual leaves exhibiting necrotic lesions and/or yellowing, caused by the insect pest coffee leaf miner (CLM), and two fungal diseases: soybean rust (SBR) and wheat tan spot (WTS). All images were manually annotated for three classes: leaf background (B), healthy leaf (H) and injured leaf (I). Precision, recall, and Intersection over Union (IoU) metrics in the test image set were the highest for B, followed by H and I classes, regardless of the architecture. When the pixel-level predictions were used to calculate percent severity, Feature Pyramid Network (FPN), Unet and DeepLabv3+ (Xception) performed the best among the architectures: concordance coefficients were greater than 0.95, 0.96 and 0.98 for CLM, SBR and WTS datasets, respectively, when confronting predictions with the annotated severity. The other three architectures tended to misclassify healthy pixels as injured, leading to overestimation of severity. Results highlight the value of a CNN-based automatic segmentation method to determine the severity on images of foliar diseases obtained under challenging conditions of brightness and background. The accuracy levels of the severity estimated by the FPN, Unet and DeepLabv3 + (Xception) were similar to those obtained by a standard commercial software, which requires adjustment of segmentation parameters and removal of the complex background of the images, tasks that slow down the process. MenosColour-thresholding digital imaging methods are generally accurate for measuring the percentage of foliar area affected by disease or pests (severity), but they perform poorly when scene illumination and background are not uniform. In this study, six convolutional neural network (CNN) architectures were trained for semantic segmentation in images of individual leaves exhibiting necrotic lesions and/or yellowing, caused by the insect pest coffee leaf miner (CLM), and two fungal diseases: soybean rust (SBR) and wheat tan spot (WTS). All images were manually annotated for three classes: leaf background (B), healthy leaf (H) and injured leaf (I). Precision, recall, and Intersection over Union (IoU) metrics in the test image set were the highest for B, followed by H and I classes, regardless of the architecture. When the pixel-level predictions were used to calculate percent severity, Feature Pyramid Network (FPN), Unet and DeepLabv3+ (Xception) performed the best among the architectures: concordance coefficients were greater than 0.95, 0.96 and 0.98 for CLM, SBR and WTS datasets, respectively, when confronting predictions with the annotated severity. The other three architectures tended to misclassify healthy pixels as injured, leading to overestimation of severity. Results highlight the value of a CNN-based automatic segmentation method to determine the severity on images of foliar diseases obtained under challenging conditions of brightness and background. The accuracy levels ... Mostrar Tudo |
|
Palavras-Chave: |
Aprendizado de máquina; Aprendizado profundo; Convolutional neural network; Fitopatometria; Image segmentation; Inteligência artificial; Machine learning; Phytopathometry; Rede neural convolucional; Segmentação de imagem. |
|
Thesagro: |
Doença de Planta. |
|
Thesaurus Nal: |
Artificial intelligence; Neural networks; Plant diseases and disorders. |
|
Categoria do assunto: |
-- |
|
URL: |
https://ainfo.cnptia.embrapa.br/digital/bitstream/item/225945/1/AP-Predictive-models-Forests-2021.pdf
|
|
Marc: |
LEADER 02980naa a2200361 a 4500
001 2134326
005 2021-09-14
008 2021 bl uuuu u00u1 u #d
024 7 $ahttps://doi.org/10.1016/j.biosystemseng.2021.08.011$2DOI
100 1 $aGONÇALVES, J. P.
245 $aDeep learning architectures for semantic segmentation and automatic estimation of severity of foliar symptoms caused by diseases or pests.$h[electronic resource]
260 $c2021
520 $aColour-thresholding digital imaging methods are generally accurate for measuring the percentage of foliar area affected by disease or pests (severity), but they perform poorly when scene illumination and background are not uniform. In this study, six convolutional neural network (CNN) architectures were trained for semantic segmentation in images of individual leaves exhibiting necrotic lesions and/or yellowing, caused by the insect pest coffee leaf miner (CLM), and two fungal diseases: soybean rust (SBR) and wheat tan spot (WTS). All images were manually annotated for three classes: leaf background (B), healthy leaf (H) and injured leaf (I). Precision, recall, and Intersection over Union (IoU) metrics in the test image set were the highest for B, followed by H and I classes, regardless of the architecture. When the pixel-level predictions were used to calculate percent severity, Feature Pyramid Network (FPN), Unet and DeepLabv3+ (Xception) performed the best among the architectures: concordance coefficients were greater than 0.95, 0.96 and 0.98 for CLM, SBR and WTS datasets, respectively, when confronting predictions with the annotated severity. The other three architectures tended to misclassify healthy pixels as injured, leading to overestimation of severity. Results highlight the value of a CNN-based automatic segmentation method to determine the severity on images of foliar diseases obtained under challenging conditions of brightness and background. The accuracy levels of the severity estimated by the FPN, Unet and DeepLabv3 + (Xception) were similar to those obtained by a standard commercial software, which requires adjustment of segmentation parameters and removal of the complex background of the images, tasks that slow down the process.
650 $aArtificial intelligence
650 $aNeural networks
650 $aPlant diseases and disorders
650 $aDoença de Planta
653 $aAprendizado de máquina
653 $aAprendizado profundo
653 $aConvolutional neural network
653 $aFitopatometria
653 $aImage segmentation
653 $aInteligência artificial
653 $aMachine learning
653 $aPhytopathometry
653 $aRede neural convolucional
653 $aSegmentação de imagem
700 1 $aPINTO, F. A. C.
700 1 $aQUEIROZ, D. M.
700 1 $aVILLAR, F. M. M.
700 1 $aBARBEDO, J. G. A.
700 1 $aDEL PONTE, E. M.
773 $tBiosystems Engineering$gv. 210, p. 129-142, Oct. 2021.
Download
Esconder MarcMostrar Marc Completo |
|
Registro original: |
Embrapa Agricultura Digital (CNPTIA) |
|
|
Biblioteca |
ID |
Origem |
Tipo/Formato |
Classificação |
Cutter |
Registro |
Volume |
Status |
URL |
Voltar
|
|
|
| Registros recuperados : 2 | |
| 2. |  | ANTUNES, A. C.; MONTANARIN, A.; GRÄBIN, D. M.; MONTEIRO, E. C. dos S. M.; PINHO, F. F. de; ALVARENGA, G. C.; AHUMADA, J.; WALLACE, R. B.; RAMALHO, E. E.; BARNETT, A. P. A.; BAGER, A.; LOPES, A. M. C.; KEUROGHLIAN, A.; GIROUX, A.; HERRERA, A. M.; CORREA, A. P. de A.; MEIGA, A. Y.; JÁCOMO, A. T. de A.; BARBAN, A. de B.; ANTUNES, A.; COELHO, A. G. de A.; CAMILO, A. R.; NUNES, A. V.; GOMES, A. C. dos S. M.; ZANZINI, A. C. da S.; CASTRO, A. B.; DESBIEZ, A. L. J.; FIGUEIREDO, A.; THOISY, B. de; GAUZENS, B.; OLIVEIRA, B. T.; LIMA, C. A. de; PERES, C. A.; DURIGAN, C. C.; BROCARDO, C. R.; ROSA, C. A.; ZÁRATE CASTAÑEDA, C.; MONTEZA MORENO, C. M.; CARNICER, C.; TRINCA, C. T.; POLLI, D. J.; FERRAZ, D. da S.; LANE, D. F.; ROCHA, D. G. da; BARCELOS, D. C.; AUZ, D.; ROSA, D. C. P.; SILVA, D. A.; SILVÉRIO, D. V.; EATON, D. P.; NAKANO OLIVEIRA, E.; VENTICINQUE, E.; JUNIOR, E. C.; MENDONÇA, E. N.; VIEIRA, E. M.; ISASI CATALÁ, E.; FISCHER, E.; CASTRO, E. P.; OLIVEIRA, E. G.; MELO, F. R. de; MUNIZ, F. de L.; ROHE, F.; BACCARO, F. B.; MICHALSKI, F.; PAIM, F. P.; SANTOS, F.; ANAGUANO, F.; PALMEIRA, F. B. L.; REIS, F. da S.; AGUIAR SILVA, F. H.; BATISTA, G. de A. B.; ZAPATA RÍOS, G.; FORERO MEDINA, G.; NETO, G. de S. F.; ALVES, G. B.; AYALA, G.; PEDERSOLI, G. H. P.; EL BIZRI, HANI R.; PRADO, H. A.; MOZERLE, H. B.; COSTA, H. C. M.; LIMA, I. J.; PALACIOS, J.; ASSIS, J. de R.; BOUBLI, J. P.; METZGER, J. P.; TEIXEIRA, J. V.; MIRANDA, J. M. D.; POLISAR, J.; SALVADOR, J.; BORGES ALMEIDA, K.; DIDIER, K.; PEREIRA, K. D. de L.; TORRALVO, K.; GAJAPERSAD, K.; SILVEIRA, L.; MAIOLI, L. U.; MARACAHIPES SANTOS, L.; VALENZUELA, L.; BENAVALLI, L.; FLETCHER, L.; PAOLUCCI, L. N.; ZANZINI, L. P.; DA SILVA, L. Z.; RODRIGUES, L. C. R.; BENCHIMOL, M.; OLIVEIRA, M. A.; LIMA, M.; DA SILVA, M. B.; SANTOS JUNIOR, M. A. dos; VISCARRA, M.; COHN HAFT, M.; ABRAHAMS, M. I.; BENEDETTI, M. A.; MARMONTEL, M.; HIRT, M. R.; TÔRRES, N. M.; CRUZ JUNIOR, O. F.; ALVAREZ LOAYZA, P.; JANSEN, P.; PRIST, P. R.; BRANDO, P. M.; PERÔNICO, P. B.; LEITE, R. do N.; RABELO, R. M.; SOLLMANN, R.; BELTRÃO MENDES, R.; FERREIRA, R. A. F.; COUTINHO, R.; OLIVEIRA, R. da C.; ILHA, R.; HILÁRIO, R. R.; PIRES, R. A. P.; SAMPAIO, R.; MOREIRA, R. da S.; BOTERO ARIAS, R.; MARTINEZ, R. V.; NÓBREGA, R. A. de A.; FADINI, R. F.; MORATO, R. G.; CARNEIRO, R. L.; ALMEIDA, R. P. S.; RAMOS, R. M.; SCHAUB, R.; DORNAS, R.; CUEVA, RUBÉN; ROLIM, S.; LAURINDO, S.; ESPINOSA, S.; FERNANDES, T. N.; SANAIOTTI, T. M.; ALVIM, T. H. G.; DORNAS, TIAGO TEIXEIRA; PIÑA, T. E. N.; ANDRADE, V. L. C.; SANTIAGO, W. T. V.; MAGNUSSON, W. E.; CAMPOS, Z.; RIBEIRO, M. C. Amazonia Camtrap: a data set of mammal, bird, and reptile species recorded with camera traps in the Amazon forest. Ecology, v. 103, n. 9, p. e3738, 2022. Datar Paper.| Biblioteca(s): Embrapa Pantanal. |
|    |
| Registros recuperados : 2 | |
|
| Nenhum registro encontrado para a expressão de busca informada. |
|
|
