A CNN-LSVM Model for Imbalanced Images Identification of Wheat Leaf

Abstract

In order to improve the accuracy of convolutional neural networks (CNN) in imbalanced dataset classification, a novel hierarchical CNN-LSVM is proposed. Considering the imbalance in the number and spatial distribution of wheat leaf disease images, the improved local support vector machine (LSVM) replaces Softmax as the classifier of the model, and meanwhile a cost sensitive matrix is designed to assign the value for penalty factors in the optimized objective function of LSVM. It effectively improves the sensitivity of misclassification caused by imbalanced data. To verify the validity and practicability of CNN-LSVM, 6028 wheat leaf disease images containing 8 species are collected from planting bases in Shandong Agricultural University. Then the imbalanced and balanced standard image sets are generated by data augmentation and Borderline-Synthetic Minority Oversampling (Borderline-SMOTE). They have 36168 and 46176 images, respectively. The experimental results show that the average identification accuracies of the CNN-LSVM obtained on imbalanced and balanced standard datasets are 90.32 % and 93.68 %, respectively. And it starts to converge when the iteration times are close to 13000. CNN-LSVM has higher classification accuracy and lower iteration times, compared with CNN-Softmax, CNN-SVM, LSVM and support vector machine (SVM).

Tackling unbalanced datasets for yellow and brown rust detection in wheat

“…This highlights the complex effects of data augmentation techniques like SMOTE on model accuracy. Similarly, Su et al. (2019) reported a 3.36% improvement in model performance when comparing outcomes on imbalanced versus balanced standard datasets, underscoring the beneficial impact of data-balancing techniques.…”

“…However, given the vastness of hyperspectral datasets and their intricate processing requirements, integrating ML models with hyperspectral data for disease identification has garnered increased interest in recent years. In this sense, models such as ANN, SVM, RF, and GNB, among others, have been proposed ( Singh et al., 2016 ; Su, 2020 ). In light of these facts, hyperspectral information for disease detection has been successfully utilized.…”

Combined CNN with STARGAN for Wheat Yellow Rust Disease Classification

“…The categorization of wheat kernels [15] has been identified through the deep learning framework VGG16 and SVM. The authors [16] proposed a hybrid approach known as CNN-LSVM and compare their approach with CNN-Softmax and CNN-SVM for wheat yellow rust identification. The authors [21] compare Continuous wavelet analysis features with the raw reflectance conventional feature technique for the identification of wheat stripe rust disease.…”

“…Therefore, machine learning techniques achieve low classification accuracy as compared to deep learning techniques for wheat rust disease identification [14], [15]. hybrid models have been developed (CNN-LSVM [16] and Differential amplification CNN (DACNN) [17] for wheat rust disease identification. After the classification of crop diseases [18], [19] different severity levels of diseases are identified at a large scale.…”

Deep Learning in Skin Disease Image Recognition: A Review

“…To evaluate the performance of deep learning in skin disease image recognition, several performance indicators are used: accuracy (ACC) represents the percentage of correct prediction results in the total sample [100]; mean average precision (mAP) represents the average accuracy of all categories [101]; true positive rate (TPR), also known as sensitivity and Recall (R) [102], represents the probability of being predicted to be positive in the actual positive samples [103]; false positive rate (FPR) refers to the percentage of actual disease-free but judged to be disease-free; true negative rate (TNR) [104], also known as specificity, indicates that the actual disease-free is correctly judged to be disease-free; area under the receiver operating characteristic (ROC) curve (AUC) refers to the probability that the classifier outputs positive and negative samples, and the likelihood that the classifier outputs a positive sample is greater than that of the negative sample; ROC is the working characteristic curve of subjects, which shows the performance of classification model under all classification thresholds [105]. The specific performance indicators are shown in Table VIII.…”