Quantification of Root-Knot Nematode Infestation in Tomato Using Digital Image Analysis

Reviewer #1’s comment

Response to  Reviewer 1

1.      The manuscript compares several methods of image analysis for the morphometry and counting of root-knot nematode stage-II juveniles. This revised version of the manuscript improves over the original version. However, the methods are not well illustrated, and the authors do not discuss potential incompatibilities between measurements.

Thank you for your comments. There are no incompatibilities between measurements. Changes in units alter the slope or the value of regression coefficient but not the statistical significance of dependent variable. We have confirmed equal coefficient of determination even after change in the units of measurement.  Changing unit does not affect the regression analysis as discussed below:

https://courses.lumenlearning.com/wmopen-concepts-statistics/chapter/linear-relationships-4-of-4/

2.      I found the term “computer vision” exaggerated in conjunction with the methods used. These are basic shape analysis methods. The authors did for example not implement artificial intelligence, spectral unmixing or 3D mapping in order to improve object separation. The authors should avoid such verbal exaggeration.

Thank you for your comment. We have clarified in the revised manuscript (right from the Abstract) we have used digital image analysis method. We have removed the term artificial intelligence and spectral unmixing or 3D mapping. However, some related research studies that used computer vision have been added according to suggestion of another reviewer. Please refer to page no 3: line:121, 137-138

3.      Soil nematodes belong to the soil micro-fauna (soil animals with <100 μm body diameter), but they are not soil micro-organisms (= soil bacteria and soil fungi), line 41.

We appreciate your comment. The term micro-organism has been replaced with micro-fauna. Please refer to page no:1, line 43.

4.     Figures 2 and 3 should be combined into one figure (e.g. “image pre-processing”) with four sub-figures. 

Figures 2 and 3 have been combined. Please refer to page 7.

5.      The nematodes in Figures 4 and 5 are far too small to illustrate the analysis methods. The differences are almost invisible. The authors should provide much larger, possibly schematic drawings to clearly illustrate the underlying principle of each method.

Figures 4 and 5 have been replaced with figures showing larger sizes of nematode images. Please refer to pages 8 and 9.

6.     The “Length” measurements in Fig. 6 are not explained in the figure legend, nor in the text.

Are measurements (2), (3) and (4) in Fig. 6 actually “Widths” rather than “Lengths”?

The length and width measurements have now been explained in the text. Please refer to page:9, line:430-432.

Yes, you are right. The measurements (2), (3) and (4) is Widths rather than Lengths. cellSens software generated these labels.

7.      Figures 7 and 8 should be superimposed into one figure, so that each sub-figure enables direct comparison between manual and automatic measurements.

Figures 7 and 8 are used to illustrate the distribution of length, width, ratio, and area. The comparison between manual and automatic measurement is showed in Figures 9 and 10. Please refer to pages 14 and 15.

8.      It appears that manual measurements of body width were averaged from several measurements along the body length (line 393), whereas image analysis methods measured the width at the midpoint along the body (line 359). In strong contrast, De Man’s morphometric measure a (= body length / greatest body diameter) refers to the maximum body width at any position along the body.

We agree that  manual measurements of body width were averaged from several measurements at middle of body length because  the manual selection of the point of maximum width is very subjective and can produce statistically significant differences between the measurements.

The image analysis methods measured the width at midpoint along the body because  the mid-body width (maximum width) was measured greatest by a computer

(https://aslopubs.onlinelibrary.wiley.com/doi/pdf/10.1002/lom3.10128). This help us to find De Man’s Morphometric parameter that is body length / greatest body diameter (ratio).

9.      The authors should discuss how differences in the definition of body width could affect the correlation of W and L/W between manual and automatic measurements in their investigation. They should also discuss how their measurements compare with published data in the taxonomic literature. Their measurements may contribute to the description of Meloidogyne incognita.

The influence of body width definition in width and ratio of manual and automatic measurement has been discussed  in page 15, lines 668-670.

The manual and automated measurement of root-knot incognita (RKN) has been compared with was found in taxonomic literature. Please refer to page no:15, line 663-666.

10.   The text contains many minor typographic and English errors, for example “peer shape”, correctly: pear shape (line 86) and “De Mans”, correctly: De Man (line 91).

English error “pear shape” has been corrected, and De Mans replaced with De Man. Please refer to page no: 2, line 88 and 93.

Reviewer #2’s comment

Response to  Reviewer 2

1. The existing literature should be classified and systematically reviewed, instead of being independently introduced one-by-one.

1. The abstract is too general and not prepared objectively. It should briefly highlight the paper's novelty as what is the main problem, how has it been resolved and where the novelty lies?

The abstract is updated. The novelty of this paper is the investigation of three new methods to detect RKN. Identifying and counting RKN is a laborious and time-consuming task. Our approach facilitates the detection and counting of RKN faster than the traditional methods. Further, it provides an optimal range of ratio (length to width) to discern and count RKN.

1. The 'conclusions' are a key component of the paper. It should complement the 'abstract' and normally used by experts to value the paper's engineering content. In general, it should sum up the most important outcomes of the paper. It should simply provide critical facts and figures achieved in this paper for supporting the claims.

The most important outcome of the paper is described in the conclusion section. Please refer to Page: 17, line:714-718

1. For better readability, the authors may expand the abbreviations at every first occurrence.

The abbreviations are expanded at every first occurrence. Please refer to page: 17 , for the abbreviation table.

1. The author should provide only relevant information related to this paper and reserve more space for the proposed framework.

Thank you for your comment. The unrelated information to this paper has been removed.

1. However, the author should compare the proposed algorithm with other recent works or provide a discussion. Otherwise, it's hard for the reader to identify the novelty and contribution of this work.

The proposed algorithm performance has been compared with existing methods. Please refer to page:15, line no:659-664

1. The descriptions given in this proposed scheme are not sufficient that this manuscript only adopted a variety of existing methods to complete the experiment where there are no strong hypothesis and methodical theoretical arguments. Therefore, the reviewer considers that this paper needs more works.

 Generally, original research article  study one approach to the problem. However, we have investigated three approaches to compute length. Further, we have investigated the optimal ratio range to detect root-knot nematodes(RKN). Besides, these two samples were explored to analyse nematode size differences.

1. Key contribution and novelty has not been detailed in manuscript. Please include it in the introduction section

Thank you for your comment. We have stated the novelty. The key contribution of this work is the implementation of three different novel approaches to find the length of root-knot nematode juvenile (J2). Further, the optimal range of ratio has been observed. It detects RKN (juvenile) with deformed shapes or self-overlapped structure. Please have a look at page 4, line no: 161-167.

1. What are the limitations of the related works

Thank you for your comments. The limitations of related works are that they are confined to other nematodes such as C. elegans. To the best of our knowledge, root-knot nematodes have not been explored. Our study focused on detecting root-knot nematodes.  Please refer to page no. 3, line no. 154-158

1. Are there any limitations of this carried out study?

The limitation of this study is it is based on size to detect juvenile. The sample must be processed well to acquire the quality image and its time taking process. Further, it could not include overlapped nematodes. Please refer to the discussion section on page no. 16 and  line no. 697-702.

1. How to select and optimize the user-defined parameters in the proposed model?

Initially, user defined parameter was set based on manual measurement. Please refer to page no. 9 line no:401-412. Then automated measurement computed and analysed data of automated measurement. These measurements were set to detect and count nematodes. Please refer to page no. 9, line 414-420. Further, the optimal parameter of the ratio was found to detect even self-overlapped nematodes.

1. There are quite a few abbreviations are used in the manuscript. It is suggested to use a table to host all the frequently used abbreviations with their descriptions to improve the readability

Thank you for your feedback. We have created an abbreviation table in Appendix A: Table 1 , Please refer to page no. 17.

1. Explain the evaluation metrics and justify why those evaluation metrics are used?

Thank you for your comment. Evaluation metrics are explained in the evaluation criteria section on page no. 12-13, line no: 497-530. We have used R2, RMSE, and ANOVA.

1. Some sentences are too long to follow, it is suggested that to break them down into short but meaningful ones to make the manuscript readable.

We have broken down longer sentences. Thank you for your feedback.

1. The title is pretty deceptive and does not address the problem completely.

The title of the manuscript is changed to “Quantification of root-knot nematode infestation in tomato using digital image analysis “. Please refer to page no. 1.

1. Every time a method/formula is used for something, it needs to be justified by either (a) prior work showing the superiority of this method, or (b) by your experiments showing its advantage over prior work methods - comparison is needed, or (c) formal proof of optimality. Please consider more prior works.

Thank you for your comment. A comparison of manual counting and automated counting has been done. R2 and RMSE metrics evaluated. Further, the comparison of the three methods was done using one-way ANOVA. The p-value was less than the significance level(α). Please refer to page no.13, line no:551-552.

1. The data is not described. Proper data description should contain the number of data items, number of parameters, distribution analysis of parameters, and of the target parameter itself for classification.

The total number of images were : Hundred and ten images of JIR and JEM were captured. Further, 570 images of JIR were acquired for investigation. Please refer to  6, line 232-233.

We have analysed manual measurement data as shown in pages 10 and 11.  We have also analysed automated measurement of ratio and area. Then after analysing the distribution of ratio and area. We set ratio between 19-30 and area between 3000-7000. Further ratio investigated at different range (15-35, 12-35, 10-35, 8-35, 6-35, and 4-35). Please refer to page 10 line, no:415-420

1. The related works section is very short and no benefits from it. I suggest increasing the number of studies and add a new discussion there to show the advantage. Following studies can be considered

Thank you for this comment. We have increased the number of studies reviewed as per your suggestion.

1. A citrus fruits and leaves dataset for detection and classification of citrus diseases through machine learning.

The two studies you suggested have also been considered. Please refer to page no. 3, lines 135-136.

1. AI-Driven Framework for Recognition of Guava Plant Diseases through Machine Learning from DSLR Camera Sensor Based High Resolution Imagery.

The two studies you suggested have also been considered. Please refer to page no. 3, lines 137-138.

1. Use Anova test to record the significant difference between performance of the proposed and existing methods.

We have used one-way ANOVA test on a different method. In addition, we have used RMSE metric. Thank you for your valuable feedback. We have addressed all your comments.