RETRACTED ARTICLE: Neutrosophic Cognitive Maps (NCM) based feature selection approach for early leaf disease diagnosis

Early diagnosis of leaf ailments is the most necessary and prominent way to increase agriculture production. In this paper, a computer-aided approach for classifying the ailments in plant leaf is proposed using the neutrosophic logic-based feature selection algorithm. Feature selection leads to bett...

Full description

Saved in:
Bibliographic Details
Published inJournal of ambient intelligence and humanized computing Vol. 12; no. 5; pp. 5627 - 5638
Main Authors Shadrach, Finney Daniel, Kandasamy, Gunavathi
Format Journal Article
LanguageEnglish
Published Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2021
Springer Nature B.V
Subjects
Accuracy
Algorithms
Artificial Intelligence
Classification
Cognitive maps
Cognitive models
Computational Intelligence
Crop diseases
Datasets
Diagnosis
Engineering
Feature selection
Genetic algorithms
Leaves
Logic
Mobile computing
Original Research
Plant diseases
Robotics and Automation
User Interfaces and Human Computer Interaction
Neutrosophic cognitive maps
Feature selection
Neural network
GLCM features
Leaf disease
Online AccessGet full text

Cover

More Information
Summary:Early diagnosis of leaf ailments is the most necessary and prominent way to increase agriculture production. In this paper, a computer-aided approach for classifying the ailments in plant leaf is proposed using the neutrosophic logic-based feature selection algorithm. Feature selection leads to better learning performance and lowers computational cost by choosing a small subset of features by eliminating noisy and redundant features thereby acting as a dimensionality reduction technique. Leaf disease classification is similar to other classification problems but varies significantly in the features that contribute to classification. In the proposed method, Neutrosophic Cognitive Maps (NCM) is used to select the best subsets from GLCM and statistical features that can effectively characterize the leaf ailments. Eight existing state-of-the-art feature selection techniques are compared with the proposed method in order to prove the ability of the proposed method on publicly available images from the PlantVillage repository. Further, the leaf diagnosis can be incorporated in a mobile computing system if needed using appropriate methods thereby enabling user-friendliness. The proposed feature selection method provides an overall classification accuracy of 99.8% while selecting just 11 features for leaf disease diagnosis
Bibliography:retraction
ISSN:1868-5137
1868-5145
DOI:10.1007/s12652-020-02070-3