Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Architecture of the CNN-BiLSTM model for leaf disease detection and classification.

More »

Fig 1 Expand

Fig 2.

Sample plant leaf images from PV database.

More »

Fig 2 Expand

Fig 3.

Block diagram of the BiLSTM architecture.

More »

Fig 3 Expand

Table 1.

Summary of the existing works.

More »

Table 1 Expand

Table 2.

Crop disease recognition using CNNs.

More »

Table 2 Expand

Fig 4.

Accuracy on crop diseases using DL models.

More »

Fig 4 Expand

Fig 5.

Dimensionality reduction/feature extraction.

More »

Fig 5 Expand

Table 3.

Crop disease recognition using CNNs.

More »

Table 3 Expand

Fig 6.

Proposed model for CNN-BiLSTM architecture.

More »

Fig 6 Expand

Fig 7.

Schematic diagram of proposed methodology.

More »

Fig 7 Expand

Table 4.

Parameter settings.

More »

Table 4 Expand

Fig 8.

Model accuracy and loss over epochs.

More »

Fig 8 Expand

Fig 9.

Training vs. validation loss for different learning rates.

More »

Fig 9 Expand

Fig 10.

Grad-CAM visualization.

More »

Fig 10 Expand

Fig 11.

Grad-CAM visualization.

More »

Fig 11 Expand

Fig 12.

t-SNE feature embedding.

More »

Fig 12 Expand

Fig 13.

Feature importance/activation maps.

More »

Fig 13 Expand

Fig 14.

Various pepper leaf disease (a, b, c, d) and Maize leaf disease (e, f, g, h) images.

More »

Fig 14 Expand

Table 5.

Disease classes with assigned numbers and image counts.

More »

Table 5 Expand

Table 6.

Disease classes with assigned numbers and image counts with training and testing.

More »

Table 6 Expand

Fig 15.

Example photos from the collection (top row represents pepper and bottom row represents maize).

More »

Fig 15 Expand

Fig 16.

Qualitative Representation (a) input image, (b) Resized image, (c) Pre-processed image and (d) Segmented image.

More »

Fig 16 Expand

Table 7.

Ablation study on various ImageJ image enhancement filters.

More »

Table 7 Expand

Table 8.

Ablation study: effect of transformer layers, activation functions, pooling layers, and stride size on leaf disease classification.

More »

Table 8 Expand

Table 9.

Ablation study effect of transformer layers, activation functions, pooling layers, and stride size on leaf disease classification.

More »

Table 9 Expand

Table 10.

Ablation study on changing kernel size, pooling layer kernel size, loss function, batch size.

More »

Table 10 Expand

Table 11.

Ablation study: effect of optimizer, learning rate, and image size on model performance.

More »

Table 11 Expand

Table 12.

Ablation studies clearly indicate that model architecture.

More »

Table 12 Expand

Table 13.

Configuration of proposed CNN-BiLSTM architecture after ablation study.

More »

Table 13 Expand

Fig 17.

Improvement in test accuracy over 11 ablation studies.

More »

Fig 17 Expand

Fig 18.

Performance with standard errors.

More »

Fig 18 Expand

Fig 19.

Comparison of suggested approach with existing works.

More »

Fig 19 Expand

Table 14.

State-of-the-art works performance comparison with proposed model.

More »

Table 14 Expand

Table 15.

Comparison of performance analysis of existing methodologies with the proposed methodology.

More »

Table 15 Expand

Fig 20.

Training set class distribution.

More »

Fig 20 Expand

Fig 21.

Test set class distribution.

More »

Fig 21 Expand

Fig 22.

BS found in test dataset.

More »

Fig 22 Expand

Fig 23.

BS and HP unaffected region.

More »

Fig 23 Expand

Fig 24.

Bacterial Spot and unaffected region identified from a Google picture.

More »

Fig 24 Expand

Fig 25.

Accuracy levels for feature extraction.

More »

Fig 25 Expand

Fig 26.

Accuracy levels for feature extraction.

More »

Fig 26 Expand

Fig 27.

Time levels for feature extraction.

More »

Fig 27 Expand

Fig 28.

Time levels for feature labelling.

More »

Fig 28 Expand

Fig 29.

Accuracy levels for feature labelling.

More »

Fig 29 Expand

Fig 30.

Feature dimensionality reduction time levels.

More »

Fig 30 Expand

Fig 31.

Accuracy levels for feature dimensionality reduction.

More »

Fig 31 Expand

Fig 32.

Confusion matrix for different models.

More »

Fig 32 Expand

Fig 33.

F1-score levels for feature dimensionality reduction.

More »

Fig 33 Expand

Fig 34.

Recall levels for features dimensionality reduction.

More »

Fig 34 Expand

Fig 35.

Precision levels for features dimensionality reduction.

More »

Fig 35 Expand

Fig 36.

Precision, Recall, F1-score comparisons of proposed model with existing models.

More »

Fig 36 Expand

Fig 37.

Accuracy comparison of the proposed model to existing models.

More »

Fig 37 Expand

Fig 38.

Testing accuracies and training parameters with the suggested work.

More »

Fig 38 Expand

Table 16.

More »

Table 16 Expand

Table 17.

Comparison of state-of-the-artworks test accuracies and training parameters with the suggested work.

More »

Table 17 Expand