Next Article in Journal
Hydrogen Sulfide Mitigates Manganese-Induced Toxicity in Malus hupehensis Plants by Regulating Osmoregulation, Antioxidant Defense, Mineral Homeostasis, and Glutathione Ascorbate Cycle
Previous Article in Journal
Evaluation of Statistical Models of NDVI and Agronomic Variables in a Protected Agriculture System
 
 
Article
Peer-Review Record

Effective Tomato Spotted Wilt Virus Resistance Assessment Using Non-Destructive Imaging and Machine Learning

Horticulturae 2025, 11(2), 132; https://doi.org/10.3390/horticulturae11020132
by Sang Gyu Kim 1,†, Sang-Deok Lee 1,†, Woo-Moon Lee 1, Hyo-Bong Jeong 2, Nari Yu 1, Oak-Jin Lee 1 and Hye-Eun Lee 1,*
Reviewer 1: Anonymous
Reviewer 2:
Reviewer 3: Anonymous
Horticulturae 2025, 11(2), 132; https://doi.org/10.3390/horticulturae11020132
Submission received: 18 December 2024 / Revised: 21 January 2025 / Accepted: 24 January 2025 / Published: 26 January 2025
(This article belongs to the Section Plant Pathology and Disease Management (PPDM))

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 I am submitting my review of the manuscript entitled "Reassessment strategy of TSWV mutation resistance in pepper lines using open software," which was submitted to Horticulturae.

I hope my comments help the authors improve the manuscript for publication.
The study proposes a non-destructive evaluation technique for Tomato Spotted Wilt Virus (TSWV) using an open software platform based on image processing and machine learning.

1.      The Introduction needs to more clearly highlight the unique contribution of the article. Why is it important? What differentiates it from existing research? What gap in the literature and practice does this study fill? Addressing these questions is crucial to justify the article and its publication.

2.      The figures of Orange are not clear for evaluation. I recommend submitting them in better quality or at a larger size for review.

3.      It is necessary to explain why the authors chose the specific methods used. Why machine learning, and why the Support Vector Machine (SVM), Logistic Regression (LR), and Neural Network (NN) models? Why not other methods?

4.      Justify why the Orange software was chosen.

5.      Explain what techniques were used to reduce bias in the analysis.

6.      The discussion section needs to be presented in conjunction with the results from other authors. Compare your results with those of other studies.

7.      Present the limitations of the study.

8.      Provide a brief conclusion with suggestions for future studies in the field.

9. General comments:
The quality of the figures needs to be improved.
A thorough spelling and grammar review should be conducted throughout the article.

Author Response

  1. The Introduction needs to more clearly highlight the unique contribution of the article. Why is it important? What differentiates it from existing research? What gap in the literature and practice does this study fill? Addressing these questions is crucial to justify the article and its publication.

>> We have added the differences and contributions from previous studies as follows (Lines 30-36).

Recently, agricultural research based on computer vision has attracted attention. Computer vision-based automated research to optimize basil seed viability evaluation [1], computer vision to automate seedling counting in horticulture [2], and an automated phenotypic parameter measurement model were proposed [3]. However, using these technologies to plant diseases and viruses is a research topic [4,5]. Recent advances in hyperspectral imaging for non-destructive testing of plant diseases and viruses are costly and labor-intensive [6]. This study aims to bridge this gap based on RGB imaging.

 

  1. The figures of Orange are not clear for evaluation. I recommend submitting them in better quality or at a larger size for review.

>> We've highlighted and changed the important parts of the figures 9 and 10 as follows.

 

Figure 9. Model development and validation based on open software Orange3-3.36.

 

Figure 10. Prediction performance evaluation based on open software Orange3-3.36.

 

  1. It is necessary to explain why the authors chose the specific methods used. Why machine learning, and why the Support Vector Machine (SVM), Logistic Regression (LR), and Neural Network (NN) models? Why not other methods?

>> The comments have been reflected as follows (Lines 93-98).

Early detection of plant diseases is necessary to improve the marketability of crops. Early diagnosis of plant diseases requires detailed investigation by experts, but it is time-consuming and expensive. Recently, computer vision has been in the spotlight as a means to replace this. Based on vision, a process of classifying image data is required to identify plant diseases. SVR, LR, and NN are classification algorithms with proven performance. Selecting algorithms with excellent performance is crucial for identifying plant diseases.

 

 

  1. Justify why the Orange software was chosen.

>> For agricultural applications of computer vision, it is important to have a method that is accessible to non-specialists. Orange Software has been selected as it meets this need effectively (Lines 190-192).

2.5.3. Machine learning

Machine learning data are images taken at 15dpi and 28dpi. From three rounds of image capture, randomly select 400 images each from TSWV-resistant and susceptible plants, totaling 800 images. Data augmentation technology improves model performance [3,4,19]. For simple geometric transformation, right-rotated, left-rotated, and vertically flipped images of the original photos were collected. Simulations for resistance determination are conducted by Orange3-3.36 software, employing models such as Support Vector Machine (SVM), Logistic Regression (LR), and Neural Network (NN). SVM is useful for finding complex boundaries, making it suitable for classification and regression, but it is computationally expensive for large datasets (Figure 8, Figure 9, Figure 10). For agricultural applications of computer vision, it is important to have a method that is accessible to non-specialists. Orange Software has been selected as it meets this need effectively.

 

  1. Explain what techniques were used to reduce bias in the analysis.

>> Machine learning is susceptible to data bias. To mitigate this issue, we collected image data in three separate rounds and randomly selected images for training (Lines 182-184).

Machine learning data are images taken at 15dpi and 28dpi. From three rounds of image capture, randomly select 400 images each from TSWV-resistant and susceptible plants, totaling 800 images. Data augmentation technology improves model performance [3,4,19]. For simple geometric transformation, right-rotated, left-rotated, and vertically flipped images of the original photos were collected. Simulations for resistance determination are conducted by Orange3-3.36 software, employing models such as Support Vector Machine (SVM), Logistic Regression (LR), and Neural Network (NN). SVM is useful for finding complex boundaries, making it suitable for classification and regression, but it is computationally expensive for large datasets (Figure 8, Figure 9, Figure 10). For agricultural applications of computer vision, it is important to have a method that is accessible to non-specialists. Orange Software has been selected as it meets this need effectively.

 

  1. The discussion section needs to be presented in conjunction with the results from other authors. Compare your results with those of other studies.

>> Performance comparison results have been added as follows (lines 325-332).

4.3. Performance comparison

Hyperspectral images are an effective tool for analyzing crop images based on machine learning [6]. The latest hyperspectral image-based crop disease detection accuracies have been reported to range from 59% to 79.2%. The accuracy level of crop disease detection can vary greatly depending on the algorithm used. These research results were similar to the accuracy level variability of the three algorithms used in this study. However, hyperspectral imaging has high hardware costs and complex data processing. RGB-based machine learning may provide a more efficient solution for identifying crop diseases.

 

  1. Present the limitations of the study.

>> The limitations of this study were expressed as follows (Lines 335-342).

  1. Conclusions

It is necessary to use the proposed methods efficiently based on research results. All three methods can, on average, serve as alternatives to address or predict the results of other methods. However, causal analysis methods that include standard deviation still lack generalization. The proposed NN model is for diagnosing the severity of virus variants ex post facto or selecting virus-resistant varieties through variant reassessment, but the spread of variants is not generalized. Developing advanced and statistically generalized techniques that can accurately assess both aspects simultaneously remains a research challenge.

 

 

  1. Provide a brief conclusion with suggestions for future studies in the field.

>> Please refer to answer 7. It has been done.

 

  1. General comments:

The quality of the figures needs to be improved.

A thorough spelling and grammar review should be conducted throughout the article.

>> Thank you. We have improved the paper as much as possible.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The topic is interesting because it presents the use of new technologies in areas such as agriculture. Below we indicate some recommendations:

References and state of the art:

- Update the references, indicating published works from 2023 and 2024.

- Increase related works in the application of computer vision in agriculture.

Materials and Methods: It is understood that image processing is carried out together with artificial intelligence techniques. We indicate the following recommendations:

- It must indicate how the image database creation process is carried out.

- How is the training process

- How is the performance measurement process.

- We are working with 3 artificial intelligence techniques.

It is recommended to make the comparison using the sensitivity and specificity measures of each technique.

Discussions: Make the comparison of your results with similar works.

Conclusions: The writing of the conclusions is not evident.

 

Author Response

The topic is interesting because it presents the use of new technologies in areas such as agriculture. Below we indicate some recommendations:

  1. References and state of the art:

1.1. Update the references, indicating published works from 2023 and 2024.

>> The references have been updated to the latest paper as follows:

Ref. 1: Altizani-Júnior, J. C., Cicero, S. M., de Lima, C. B., Alves, R. M., & Gomes-Junior, F. G. (2024). Optimizing Basil Seed Vigor Evaluations: An Automatic Approach Using Computer Vision-Based Technique. Horticulturae, 10(11), 1220.

Ref. 2: Fuentes-Peñailillo, F., Carrasco Silva, G., Pérez Guzmán, R., Burgos, I., & Ewertz, F. (2023). Automating seedling counts in horticulture using computer vision and AI. Horticulturae, 9(10), 1134.

Ref. 3: Zhang, W., Dang, L. M., Nguyen, L. Q., Alam, N., Bui, N. D., Park, H. Y., & Moon, H. (2024). Adapting the Segment Anything Model for Plant Recognition and Automated Phenotypic Parameter Measurement. Horticulturae, 10(4), 398.

Ref. 4: García-Amaro, E., Cervantes-Canales, J., García-Lamont, F., Lara-Viveros, F. M., Ruiz-Castilla, J. S., & Espejel Cabrera, J. (2024). Use of Computer Vision Techniques for Recognition of Diseases and Pests in Tomato Plants. Computación y Sistemas, 28(2), 709-723.

Ref. 5: Clohessy, J. W., Sanjel, S., O'Brien, G. K., Barocco, R., Kumar, S., Adkins, S., ... & Small, I. M. (2021). Development of a high-throughput plant disease symptom severity assessment tool using machine learning image analysis and integrated geolocation. Computers and Electronics in Agriculture, 184, 106089.

Ref. 6: García-Vera, Y. E., Polochè-Arango, A., Mendivelso-Fajardo, C. A., & Gutiérrez-Bernal, F. J. (2024). Hyperspectral image analysis and machine learning techniques for crop disease detection and identification: A review. Sustainability, 16(14), 6064.

Ref. 12: Lafrance, R., Villicaña, C., Valdéz-Torres, J. B., García-Estrada, R. S., Báez Sañudo, M. A., Esparza-Araiza, M. J., & León-Félix, J. (2024). Selection of Tomato (Solanum lycopersicum) Hybrids Resistant to Fol, TYLCV, and TSWV with Early Maturity and Good Fruit Quality. Horticulturae, 10(8), 839.

Ref. 13: Oh, B. G., Chung, J. S., Ju, H. J., Yoon, J. Y., & Baek, E. (2024). Arctium lappa is a New Natural Host of Tomato Spotted Wilt Virus in Korea. Plant Disease, 108(10), 3206.

Ref. 14: Abdisa, E., Kwon, J., Jin, G., & Kim, Y. (2024). Thrips and TSWV Occurrence in Geographically Different Open Fields Cultivating Hot Peppers. Korean journal of applied entomology, 63(1), 43-51.

Ref. 15: Shymanovich, T., Saville, A. C., Mohammad, N., Wei, Q., Rasmussen, D., Lahre, K. A., ... & Ristaino, J. B. (2024). Disease progress and detection of a California resistance-breaking strain of tomato spotted wilt virus in tomato with LAMP and CRISPR-Cas12a assays. PhytoFrontiers™, 4(1), 50-60.

Ref. 16: Juárez, I. D., Steczkowski, M. X., Chinnaiah, S., Rodriguez, A., Gadhave, K. R., & Kurouski, D. (2024). Using Raman spectroscopy for early detection of resistance-breaking strains of tomato spotted wilt orthotospovirus in tomatoes. Frontiers in Plant Science, 14, 1283399.

Ref. 17: Caruso, A. G., Ragona, A., Agrò, G., Bertacca, S., Yahyaoui, E., Galipienso, L., ... & Davino, S. (2024). Rapid detection of tomato spotted wilt virus by real-time RT-LAMP and in-field application. Journal of plant pathology, 1-16.

Ref. 19: Parisi, M., Alioto, D., & Tripodi, P. (2020). Overview of biotic stresses in pepper (Capsicum spp.): Sources of genetic resistance, molecular breeding and genomics. International Journal of Molecular Sciences, 21(7), 2587.

Ref. 28: Orecchio, C., Botto, C. S., Alladio, E., D'Errico, C., Vincenti, M., & Noris, E. (2025). Non-invasive and early detection of tomato spotted wilt virus infection in tomato plants using a hand-held Raman spectrometer and machine learning modelling. Plant Stress, 100732.

 

1.2. Increase related works in the application of computer vision in agriculture.

>> Research related to agricultural computer vision has increased as follows (line30-36).

Recently, agricultural research based on computer vision has attracted attention. Computer vision-based automated research to optimize basil seed viability evaluation [26], computer vision to automate seedling counting in horticulture [27], and an automated phenotypic parameter measurement model were proposed [28]. However, using these technologies to plant diseases and viruses is a research topic [29,30]. Recent advances in hyperspectral imaging for non-destructive testing of plant diseases and viruses are costly and labor-intensive [31]. This study aims to bridge this gap based on RGB imaging.

  1. Materials and Methods: It is understood that image processing is carried out together with artificial intelligence techniques. We indicate the following recommendations:

2.1. It must indicate how the image database creation process is carried out.

>> The process of generating image data has been added as follows (Lines 48-51).

The image data generation process follows: 1. Turn on the light. 2. Check the illuminance inside the box using a spectrometer. 3. Acquire a color checker image to check the quality of the image data. The color checker is for calibration. 4. Acquire an image of the plant object.

2.2. How is the training process

>> The learning process uses the Orange software tool, where the acquired images are corrected, vectorized, and saved as .csv files. This methodology is discussed in the paper as follows (Lines 172-180).

Figure 7 shows the preprocessing for machine learning on the collected image data. The original image data has a three-dimensional tensor structure, which does not directly apply to the machine learning library used in this study. We converted the 3-dimensional tensor data into 1-dimensional vector data, normalized it, saved it in CSV file format, and applied it to the machine learning model.

 

Figure 7. Image preprocessing where  is a tensor,  is a vector,  is a scalar for a specific feature ,  is a maximum value of features, and  is a minimum value of features.

 

2.3. How is the performance measurement process.

>> Performance was measured quantitatively using the following performance measurement metrics (Line 218).

2.5.4. Metrics

 

(1)

 

(2)

 

(3)

 

(4)

 

(5)

 

where TP is true and positive, FP is false and Positive, FN is false and negative, and TN is true and negative.

  1. We are working with 3 artificial intelligence techniques.

It is recommended to make the comparison using the sensitivity and specificity measures of each technique.

>> As you mentioned, we agree that it is necessary to check the tradeoff characteristics of specificity vs. sensitivity both statistically and pathologically. However, our tool in this study provided precision and recall metrics. In further studies, we plan to use specificity-sensitivity as a measurement tool. Thank you.

  1. Discussions: Make the comparison of your results with similar works.

>> Performance comparison results have been added as follows (Line 325-332).

4.3. Performance comparison

Hyperspectral images are an effective tool for analyzing crop images based on machine learning [6]. The latest hyperspectral image-based crop disease detection accuracies have been reported to range from 59% to 79.2%. The accuracy level of crop disease detection can vary greatly depending on the algorithm used. These research results were similar to the accuracy level variability of the three algorithms used in this study. However, hyperspectral imaging has high hardware costs and complex data processing. RGB-based machine learning may provide a more efficient solution for identifying crop diseases.

 

  1. Conclusions: The writing of the conclusions is not evident.

>> Thank you.

  1. Conclusions

It is necessary to use the proposed methods efficiently based on research results. All three methods can, on average, serve as alternatives to address or predict the results of other methods. However, causal analysis methods that include standard deviation still lack generalization. The proposed NN model is for diagnosing the severity of virus variants ex post facto or selecting virus-resistant varieties through variant reassessment, but the spread of variants is not generalized. Developing advanced and statistically generalized techniques that can accurately assess both aspects simultaneously remains a research challenge.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The article addresses a key requirement for effective Tomato Spotted Wilt Virus (TSWV) resistance evaluation by non-destructive imaging and machine learning. While innovative, the change from empirical visual surveys to automated methods necessitates a greater focus on difficulties such as scalability and cost. The study's ELISA validation increases its trustworthiness, although the dataset's variety and model optimization are not adequately investigated. Furthermore, further information on imaging methods and machine learning operations is required for repeatability. The should concentrate on discussion of integrating this technology into actual agricultural systems and overcoming deployment hurdles. Moreover, the following questions and suggestion can further enhance the quality of the manuscript.

Specific suggestions:

Remove abbreviation from the title, give full form.

The title does not represent the content of the study, kindly revise it.

Highlight the objectives at the end of the introduction sections as a, b and c

How red peppers test support the authentication of the results? What are its results?

How many plants used to acquire the image data?

Fig.8 – 10: Please clear the figure, currently it is absurd. To provide only interphase is not appropriate. All the results need proper representation.

The discussion section is very poor.

Integrate the results of the disease indices or levels with image data. Make new analysis about the results in respect to disease severity.

Overall the results presentation is very poor.

There is not any analysis that can show or validate the quality of the data.

Also enlist the previous studies that have used same system for image analysis.

The summary findings in the discussion section are very superficial and not new, please make deep discussion on them.

 

The conclusion section is missing. 

Author Response

Comments and Suggestions for Authors

The article addresses a key requirement for effective Tomato Spotted Wilt Virus (TSWV) resistance evaluation by non-destructive imaging and machine learning. While innovative, the change from empirical visual surveys to automated methods necessitates a greater focus on difficulties such as scalability and cost. The study's ELISA validation increases its trustworthiness, although the dataset's variety and model optimization are not adequately investigated. Furthermore, further information on imaging methods and machine learning operations is required for repeatability. The should concentrate on discussion of integrating this technology into actual agricultural systems and overcoming deployment hurdles. Moreover, the following questions and suggestion can further enhance the quality of the manuscript.

>> Thank you.

Specific suggestions:

  1. Remove abbreviation from the title, give full form.
  2. The title does not represent the content of the study, kindly revise it.

>> The title has been revised as follows:

Effective Tomato Spotted Wilt Virus (TSWV) Resistance Assessment Using Non-Destructive Imaging and Machine Learning

 

  1. Highlight the objectives at the end of the introduction sections as a, b and c

>> The objectives have been highlighted as follows (Lines 84-87).

By combining TSWV resistance assessment and open software-based machine learning technology, this study aims to improve a. the accuracy, b. the efficiency of diagnosis, and c. establish a pepper digital breeding system to prepare for new mutations. The overall scheme of this study is shown in Figure 1.

 

  1. How red peppers test support the authentication of the results? What are its results?

>> A total of 1,925 images were used, with 800 photos to build the model and 1,125 images for prediction, achieving a classification accuracy of 82.9% in the prediction. (Figures 5 and 10).

 

  1. How many plants used to acquire the image data?

>> A total of 25 lines were inoculated with 13 individuals each for disease screening, and the experiment was repeated three times (Lines 92-106).

 

  1. Fig.8 – 10: Please clear the figure, currently it is absurd. To provide only interphase is not appropriate. All the results need proper representation.

>> We've highlighted and changed the necessary parts of Figures 9 and 10.

  1. The discussion section is very poor.

>> Performance comparison results have been added as follows (Lines 325-332).

4.3. Performance comparison

Hyperspectral images are an effective tool for analyzing crop images based on machine learning [6]. The latest hyperspectral image-based crop disease detection accuracies have been reported to range from 59% to 79.2%. The accuracy level of crop disease detection can vary greatly depending on the algorithm used. These research results were similar to the accuracy level variability of the three algorithms used in this study. However, hyperspectral imaging has high hardware costs and complex data processing. RGB-based machine learning may provide a more efficient solution for identifying crop diseases.

 

  1. Integrate the results of the disease indices or levels with image data. Make new analysis about the results in respect to disease severity.

>> We appreciate your suggestion to integrate and analyze the results of visual investigations alongside machine learning. However, we would like to emphasize that performing such an integrated analysis is challenging under the current research conditions. Therefore, we request that you consider this difficulty when evaluating the analysis results presented in this paper.

 

  1. Overall the results presentation is very poor.

>> We have improved the paper as much as possible

 

  1. There is not any analysis that can show or validate the quality of the data.

>> We have added the following validation results from analyzing data quality (Lines 246-260).

The NN model exhibited the highest prediction accuracy with a Classification Accuracy (CA) value of 0.860, significantly outperforming the LR and SVM models, which had CA values of 0.809 and 0.653, respectively (Figure 13). The NN model also demonstrated superior performance in Area Under the ROC Curve (AUC), Precision, Recall, and F1 Score, highlighting its robustness in classifying TSWV resistance.

Based on modeling performance, the superiority of the NN model was confirmed. Next, verification of the prediction performance of the NN model was performed. As a result, in the validation set of 960 images not used for training, the NN model also achieved the highest prediction accuracy with a CA of 0.843. Using the developed model, TSWV infection was predicted for the remaining 1125 images. Among the Neural Network (NN), Logistic Regression (LR), and Support Vector Machine (SVM) models, the Neural Network model showed the highest classification accuracy at 0.829 (Figure 10).

  1. Also enlist the previous studies that have used same system for image analysis.

>> We have added and compared the results of similar studies to this study as follows (Lines 30-36).

Recently, agricultural research based on computer vision has attracted attention. Computer vision-based automated research to optimize basil seed viability evaluation [1], computer vision to automate seedling counting in horticulture [2], and an automated phenotypic parameter measurement model were proposed [3]. However, using these technologies to plant diseases and viruses is a research topic [4,5]. Recent advances in hyperspectral imaging for non-destructive testing of plant diseases and viruses are costly and labor-intensive [6]. This study aims to bridge this gap based on RGB imaging.

 

  1. The summary findings in the discussion section are very superficial and not new, please make deep discussion on them.

>> For in-depth discussion in the discussion section, we have added performance comparisons with state-of-the-art technologies as follows (Lines 328-335).

4.3. Performance comparison

Hyperspectral images are an effective tool for analyzing crop images based on ma-chine learning [6]. The latest hyperspectral image-based crop disease detection accuracies have been reported to range from 59% to 79.2%. The accuracy level of crop disease detection can vary greatly depending on the algorithm used. These research results were similar to the accuracy level variability of the three algorithms used in this study. However, hyperspectral imaging has high hardware costs and complex data processing. RGB-based machine learning may provide a more efficient solution for identifying crop diseases.

 

  1. The conclusion section is missing. 

>> Thank you.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

 Overall Recommendation

Author Response

We sincerely thank the reviewers and the editor for their constructive comments and valuable suggestions to improve the quality of our manuscript. We appreciate the overall recommendation and have thoroughly addressed all the comments provided by the reviewers.

Thank you for your constructive feedback.

Reviewer 3 Report

Comments and Suggestions for Authors

Only three modifications!

1. remove abbreviation from the title

2. Give each objective with a proper sentence description, not just one word. 

3. Revise the conclusion section professionally. 

Author Response

Dear Reviewer,

We sincerely appreciate the time and effort you have dedicated to reviewing our manuscript. Your insightful comments have significantly improved the quality of our work. Below, we provide detailed responses to each of the three requested modifications:

 

  1. remove abbreviation from the title

Response: Thank you for pointing this out. We have removed the abbreviation "TSWV" from the title to enhance clarity for a broader audience. The revised title now reads:

“Effective tomato spotted wilt virus resistance assessment using non-destructive imaging and machine learning”

 

  1. Give each objective with a proper sentence description, not just one word. 

Response: We agree that clear and detailed descriptions of the objectives enhance the overall readability and comprehension of the manuscript. The objectives have been rewritten as follows:

 

Previous version:

By combining TSWV resistance assessment and open software-based machine learning technology, this study aims to improve a. the accuracy, b. the efficiency of diag-nosis, and c. establish a pepper digital breeding system to prepare for new mutations. The overall scheme of this study is shown in Figure 1.

 

Revised version:

This study aims to integrate TSWV resistance assessment with open software-based machine learning techniques to achieve the following objectives: (a) To improve the accu-racy of TSWV resistance assessment. (b) To provide an efficient diagnostic method that complements visual inspection and ELISA by utilizing RGB imaging and machine learn-ing. (c) To propose a digital breeding approach capable of addressing new mutations. The comprehensive framework of this study is illustrated in Figure 1.

 

  1. Revise the conclusion section professionally. 

 

Response: We have thoroughly revised the conclusion section to present the findings more professionally, emphasizing the significance and potential applications of our research. The revised conclusion now reads:

 

Previous version:

It is necessary to use the proposed methods efficiently based on research results. All three methods can, on average, serve as alternatives to address or predict the results of other methods. However, causal analysis methods that include standard deviation still lack generalization. The proposed NN model is for diagnosing the severity of virus variants ex post facto or selecting virus-resistant varieties through variant reassessment, but the spread of variants is not generalized. Developing advanced and statistically generalized techniques that can accurately assess both aspects simultaneously remains a research challenge.

 

Revised version:

State-of-the-art techniques for assessing TSWV resistance in tomatoes include visual inspection, ELISA assays, and imaging and machine learning methods. The proposed imaging and machine learning-based assessment approaches hold significant potential to advance sustainable agriculture by conserving environmental resources and enhancing productivity. With the rapid progress in computational technology, the accuracy of disease resistance assessment methods is expected to improve, leading to the wider adoption of non-destructive techniques for determining disease resistance. It is necessary to use the proposed methods efficiently based on research results. All three methods can, on average, serve as alternatives to address or predict the results of other methods. However, causal analysis methods that include standard deviation still lack generalization. The proposed NN model is for diagnosing the severity of virus variants ex post facto or selecting virus-resistant varieties through variant reassessment, but the spread of variants is not generalized. Developing advanced and statistically generalized techniques that can accurately assess both aspects simultaneously remains a research challenge.

 

 

We believe these revisions enhance the clarity and impact of our manuscript.

________________________________________

We hope that these changes adequately address your concerns. Please feel free to let us know if further modifications are needed.

Thank you for your constructive feedback.

Back to TopTop