Matrib leaf classification using Deep Neural Network: An integrated image processing technique

Hanif Bhuiyan, Md Abdul Karim, Jinat Ara, Abdullah Al Omar, Saharul Islam, Faria Benta Karim, Guido Governatori

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

2 Citations (Scopus)

Abstract

Healthy farm plant leaf classification and identification is a critical food security issue. In many places of the world, it remains tough as it needs appropriate infrastructure. Combining the rising worldwide prevalence of the smartphone with current progress in computer vision through deep learning, now it is possible to diagnose inconsistency of various farm plants. In this technology era, automation can help to replace manual prevention efforts in plants by employing image processing methods. This research deployed three pre-trained deep neural models: 3DCNN, ResNet50 and MobileNet, to classify the Matrib leaf into two categories: Good Matrib leaf and Bad Matrib leaf. We employed our own Matrib leaf customized dataset for this research. Experimental results demonstrate that MobileNet outperformed other models with an accuracy of 99.99% on test data, while ResNet50 and 3DCNN followed with an accuracy of 92.67% and 72.80%.

Original languageEnglish
Title of host publication2022 IEEE Region 10 Symposium (TENSYMP)
EditorsYashwant Gupta
Place of PublicationUSA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Number of pages6
ISBN (Electronic)9781665466585
DOIs
Publication statusPublished - 2022
Externally publishedYes
EventIEEE Region 10 Symposium 2022 - Mumbai, India
Duration: 1 Jul 20223 Jul 2022
https://www.ieeebombay.org/tensymp2022/
https://ieeexplore.ieee.org/xpl/conhome/9864330/proceeding

Conference

ConferenceIEEE Region 10 Symposium 2022
Abbreviated titleTENSYMP 2022
Country/TerritoryIndia
CityMumbai
Period1/07/223/07/22
Internet address

Keywords

  • Convolution neural network
  • Deep learning
  • Image processing
  • Matrib leaf
  • MobileNet
  • Resnet50

Cite this