A mobile-based deep learning model for cassava disease diagnosis

Amanda Ramcharan, Peter McCloskey, Kelsee Baranowski, Neema Mbilinyi, Latifa Mrisho, Mathias Ndalahwa, James Legg, David Peter Hughes

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Convolutional neural network (CNN) models have the potential to improve plant disease phenotyping where the standard approach is visual diagnostics requiring specialized training. In scenarios where a CNN is deployed on mobile devices, models are presented with new challenges due to lighting and orientation. It is essential for model assessment to be conducted in real world conditions if such models are to be reliably integrated with computer vision products for plant disease phenotyping. We train a CNN object detection model to identify foliar symptoms of diseases in cassava (Manihot esculenta Crantz). We then deploy the model in a mobile app and test its performance on mobile images and video of 720 diseased leaflets in an agricultural field in Tanzania. Within each disease category we test two levels of severity of symptoms-mild and pronounced, to assess the model performance for early detection of symptoms. In both severities we see a decrease in performance for real world images and video as measured with the F-1 score. The F-1 score dropped by 32% for pronounced symptoms in real world images (the closest data to the training data) due to a decrease in model recall. If the potential of mobile CNN models are to be realized our data suggest it is crucial to consider tuning recall in order to achieve the desired performance in real world settings. In addition, the varied performance related to different input data (image or video) is an important consideration for design in real world applications.

Original languageEnglish (US)
Article number272
JournalFrontiers in Plant Science
Volume10
DOIs
StatePublished - Mar 7 2019

Fingerprint

disease diagnosis
cassava
learning
neural networks
signs and symptoms (plants)
plant diseases and disorders
phenotype
Manihot esculenta
computer vision
Tanzania
lighting
testing

All Science Journal Classification (ASJC) codes

  • Plant Science

Cite this

Ramcharan, A., McCloskey, P., Baranowski, K., Mbilinyi, N., Mrisho, L., Ndalahwa, M., ... Hughes, D. P. (2019). A mobile-based deep learning model for cassava disease diagnosis. Frontiers in Plant Science, 10, [272]. https://doi.org/10.3389/fpls.2019.00272
Ramcharan, Amanda ; McCloskey, Peter ; Baranowski, Kelsee ; Mbilinyi, Neema ; Mrisho, Latifa ; Ndalahwa, Mathias ; Legg, James ; Hughes, David Peter. / A mobile-based deep learning model for cassava disease diagnosis. In: Frontiers in Plant Science. 2019 ; Vol. 10.
@article{5ef00c0dd68d41faa3fd46cab4fc1f48,
title = "A mobile-based deep learning model for cassava disease diagnosis",
abstract = "Convolutional neural network (CNN) models have the potential to improve plant disease phenotyping where the standard approach is visual diagnostics requiring specialized training. In scenarios where a CNN is deployed on mobile devices, models are presented with new challenges due to lighting and orientation. It is essential for model assessment to be conducted in real world conditions if such models are to be reliably integrated with computer vision products for plant disease phenotyping. We train a CNN object detection model to identify foliar symptoms of diseases in cassava (Manihot esculenta Crantz). We then deploy the model in a mobile app and test its performance on mobile images and video of 720 diseased leaflets in an agricultural field in Tanzania. Within each disease category we test two levels of severity of symptoms-mild and pronounced, to assess the model performance for early detection of symptoms. In both severities we see a decrease in performance for real world images and video as measured with the F-1 score. The F-1 score dropped by 32{\%} for pronounced symptoms in real world images (the closest data to the training data) due to a decrease in model recall. If the potential of mobile CNN models are to be realized our data suggest it is crucial to consider tuning recall in order to achieve the desired performance in real world settings. In addition, the varied performance related to different input data (image or video) is an important consideration for design in real world applications.",
author = "Amanda Ramcharan and Peter McCloskey and Kelsee Baranowski and Neema Mbilinyi and Latifa Mrisho and Mathias Ndalahwa and James Legg and Hughes, {David Peter}",
year = "2019",
month = "3",
day = "7",
doi = "10.3389/fpls.2019.00272",
language = "English (US)",
volume = "10",
journal = "Frontiers in Plant Science",
issn = "1664-462X",
publisher = "Frontiers Media S. A.",

}

Ramcharan, A, McCloskey, P, Baranowski, K, Mbilinyi, N, Mrisho, L, Ndalahwa, M, Legg, J & Hughes, DP 2019, 'A mobile-based deep learning model for cassava disease diagnosis', Frontiers in Plant Science, vol. 10, 272. https://doi.org/10.3389/fpls.2019.00272

A mobile-based deep learning model for cassava disease diagnosis. / Ramcharan, Amanda; McCloskey, Peter; Baranowski, Kelsee; Mbilinyi, Neema; Mrisho, Latifa; Ndalahwa, Mathias; Legg, James; Hughes, David Peter.

In: Frontiers in Plant Science, Vol. 10, 272, 07.03.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A mobile-based deep learning model for cassava disease diagnosis

AU - Ramcharan, Amanda

AU - McCloskey, Peter

AU - Baranowski, Kelsee

AU - Mbilinyi, Neema

AU - Mrisho, Latifa

AU - Ndalahwa, Mathias

AU - Legg, James

AU - Hughes, David Peter

PY - 2019/3/7

Y1 - 2019/3/7

N2 - Convolutional neural network (CNN) models have the potential to improve plant disease phenotyping where the standard approach is visual diagnostics requiring specialized training. In scenarios where a CNN is deployed on mobile devices, models are presented with new challenges due to lighting and orientation. It is essential for model assessment to be conducted in real world conditions if such models are to be reliably integrated with computer vision products for plant disease phenotyping. We train a CNN object detection model to identify foliar symptoms of diseases in cassava (Manihot esculenta Crantz). We then deploy the model in a mobile app and test its performance on mobile images and video of 720 diseased leaflets in an agricultural field in Tanzania. Within each disease category we test two levels of severity of symptoms-mild and pronounced, to assess the model performance for early detection of symptoms. In both severities we see a decrease in performance for real world images and video as measured with the F-1 score. The F-1 score dropped by 32% for pronounced symptoms in real world images (the closest data to the training data) due to a decrease in model recall. If the potential of mobile CNN models are to be realized our data suggest it is crucial to consider tuning recall in order to achieve the desired performance in real world settings. In addition, the varied performance related to different input data (image or video) is an important consideration for design in real world applications.

AB - Convolutional neural network (CNN) models have the potential to improve plant disease phenotyping where the standard approach is visual diagnostics requiring specialized training. In scenarios where a CNN is deployed on mobile devices, models are presented with new challenges due to lighting and orientation. It is essential for model assessment to be conducted in real world conditions if such models are to be reliably integrated with computer vision products for plant disease phenotyping. We train a CNN object detection model to identify foliar symptoms of diseases in cassava (Manihot esculenta Crantz). We then deploy the model in a mobile app and test its performance on mobile images and video of 720 diseased leaflets in an agricultural field in Tanzania. Within each disease category we test two levels of severity of symptoms-mild and pronounced, to assess the model performance for early detection of symptoms. In both severities we see a decrease in performance for real world images and video as measured with the F-1 score. The F-1 score dropped by 32% for pronounced symptoms in real world images (the closest data to the training data) due to a decrease in model recall. If the potential of mobile CNN models are to be realized our data suggest it is crucial to consider tuning recall in order to achieve the desired performance in real world settings. In addition, the varied performance related to different input data (image or video) is an important consideration for design in real world applications.

UR - http://www.scopus.com/inward/record.url?scp=85064205542&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064205542&partnerID=8YFLogxK

U2 - 10.3389/fpls.2019.00272

DO - 10.3389/fpls.2019.00272

M3 - Article

VL - 10

JO - Frontiers in Plant Science

JF - Frontiers in Plant Science

SN - 1664-462X

M1 - 272

ER -

Ramcharan A, McCloskey P, Baranowski K, Mbilinyi N, Mrisho L, Ndalahwa M et al. A mobile-based deep learning model for cassava disease diagnosis. Frontiers in Plant Science. 2019 Mar 7;10. 272. https://doi.org/10.3389/fpls.2019.00272