Aggregating Different Scales of Attention on Feature Variants for Tomato Leaf Disease Diagnosis from Image Data: A Transformer Driven Study

Abstract

Tomato leaf diseases can incur significant financial damage by having adverse impacts on crops and, consequently, they are a major concern for tomato growers all over the world. The diseases may come in a variety of forms, caused by environmental stress and various pathogens. An automated approach to detect leaf disease from images would assist farmers to take effective control measures quickly and affordably. Therefore, the proposed study aims to analyze the effects of transformer-based approaches that aggregate different scales of attention on variants of features for the classification of tomato leaf diseases from image data. Four state-of-the-art transformer-based models, namely, External Attention Transformer (EANet), Multi-Axis Vision Transformer (MaxViT), Compact Convolutional Transformers (CCT), and Pyramid Vision Transformer (PVT), are trained and tested on a multiclass tomato disease dataset. The result analysis showcases that MaxViT comfortably outperforms the other three transformer models with 97% overall accuracy, as opposed to the 89% accuracy achieved by EANet, 91% by CCT, and 93% by PVT. MaxViT also achieves a smoother learning curve compared to the other transformers. Afterwards, we further verified the legitimacy of the results on another relatively smaller dataset. Overall, the exhaustive empirical analysis presented in the paper proves that the MaxViT architecture is the most effective transformer model to classify tomato leaf disease, providing the availability of powerful hardware to incorporate the model.

Keywords: CCT; EANet; MaxViT; PVT; attention; tomato leaf disease; transformers.

Figure 1

Sample of the first dataset used in the study.

Figure 2

Sample of the second dataset used in the study.

Figure 3

External-attention for EANet model [31].

Figure 4

MaxViT architecture [32]. Note that this model uses both local and global mechanisms via the maxViT block.

Figure 5

MaxViT block to capture both local and global attention information [32].

Figure 6

Compact Convolutional Transformer (CCT) used as classifier.

Figure 7

PVT architecture [34].

Figure 8

Summary of the methodology.

Figure 9

Accuracy and loss curves for the models.

Figure 10

Classification report.

Figure 11

Confusion matrices.

Figure 12

ROC for each of the models.

Figure 13

Inconsistent pattern in the Septoria Leaf Spot class.

Figure 14

Classification report.

Figure 15

Accuracy and loss curves for the models.