Detection of Hardening in Mangosteens Using near-Infrared Hyperspectral Imaging

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

In my opinion, the manuscript is generally well-written and informative. I believe it can be considered for further process, however, I would like to draw your attention to the following issues:

1.      From the text (Materials and Methods section) follows that the pedicels were not removed before scanning. Could it have any was their influence on the obtained results?

2.      Line 83: The KMITL abbreviation was used. Please consider writing the Institute’s name in the whole meaning, especially since it is written only once throughout the whole text.

3.      Section 2.3 Firmness measurement: Please provide information on a number of replicates of the drop experiment for each height and on the number of undamaged mangosteens that were considered as the control group.

4.      Lines 131-133: Was the point of firmness measurement with the use of a texture analyzer placed inside the ROI? If yes, what was the influence of damage caused by texture measurements on the HSI?

5.      Figure 7 and its analysis

a.      Please declare clearly the meaning of values above each image, and introduce the unit for presented numbers.

b.      The images are predictive results for damaged mangosteens at different degrees of hardening. Please provide information if there is a possibility to connect each image with the height from which the real sample was dropped. If yes, please include the information on the height of the drop.

6.      If possible, please consider widening the analysis of mixed groups for sizes that assure statistical significance of firmness measured with a texture analyzer (for instance a mixed group of sizes 1 and 2 or 1 and 3)

I believe that considering my comments no. 1-5 can improve the comprehensiveness of the manuscript and introducing the results mentioned in my comment no. 6 would expand the understanding of the proper model selection.

Author Response

Responses to the respected editor and reviewers:

 

We would like to thank all respected reviewers for valuable comments that have helped us to improve our manuscript.  Changes have been highlighted in yellow within the manuscript.  The revised manuscript has been edited the language by the native English speaker again.  We hope you accept these corrections. Our point-by-point answers to the comments are in the following:

 

Reviewer 1

Comments and Suggestions for Authors

 

In my opinion, the manuscript is generally well-written and informative. I believe it can be considered for further process; however, I would like to draw your attention to the following issues:

Thank you very much for your positive feedback on our manuscript. We have modified the manuscript as your comments and suggestions.

1. From the text (Materials and Methods section) follows that the pedicels were not removed before scanning. Could it have any was their influence on the obtained results?

Our response to the respected reviewer: Thank you very much for your question. The pedicels did not have an influence on the obtained results because the main data for establishing the model were spectral data and texture value in ROI. This area of ROI was only on the surface of pericarp. The HSI in this study was used as a nondestructive technique. Therefore, the pedicels were not removed before scanning.

 

2. Line 83: The KMITL abbreviation was used. Please consider writing the Institute’s name in the whole meaning, especially since it is written only once throughout the whole text.

Our response to the respected reviewer: Thank you very much for your kind attention. The manuscript has been improved as your kind suggestion.

 

Line 91-93: they were cleaned and carefully packed and transported to the laboratory at King Mongkut’s Institute of Technology Ladkrabang (KMITL) in Bangkok.

 

3. Section 2.3 Firmness measurement: Please provide information on a number of replicates of the drop experiment for each height and on the number of undamaged mangosteens that were considered as the control group.

Our response to the respected reviewer: Thank you very much for your kind comment for bringing up this point. The manuscript has been improved as your kind suggestion.

 

Line 118-120: The firmness of each fruit was determined, and the number of sound mangosteens was 50 samples that were used for considering the among different sizes of mangosteen.

 

Line 130-133: The number of samples used for the dropping experiment was 560 of which 280 were from the group of size 3 and also 280 from the mixed-size group of size 3 and 4. The height of drop varied from 50 to 100 cm onto a cement floor and was randomly out onto a cement floor.

 

Line 185-186: Samples of both sound and hardened mangosteens were used for creating the predictive images based on firmness for comparison.

4. Lines 131-133: Was the point of firmness measurement with the use of a texture analyzer placed inside the ROI? If yes, what was the influence of damage caused by texture measurements on the HSI?

Our response to the respected reviewer: Thank you very much for your question.

Yes, the firmness measurement with the use of a texture analyzer placed inside the ROI. Due to the firmness value from texture measurements had the influence on the characteristic of acquired spectra from HSI. Therefore, the firmness value and the spectral data of each sample from the ROI were used as the representative data for establishing the model.

 

Line 161-164: The firmness values from the texture measurements could influence the characteristic of acquired spectra, therefore, the firmness value and the spectral data of each sample from the ROI were used as the representative data for establishing the model.

 

5. Figure 7 and its analysis
6. a. Please declare clearly the meaning of values above each image, and introduce the unit for presented numbers.

Our response to the respected reviewer: Thank you very much for your kind attention. The meaning of value above each mangosteen image was averaged firmness that was calculated from each pixel from the predictive image of each mangosteen. The manuscript has been improved as your kind suggestion.

 

Line 302-303: The averaged firmness value that was calculated from each pixel from each predictive image was presented above each mangosteen image.

 

Figure 7 has been corrected as your kind comment. Please see the Figure 7 in the attached file for the figure.

 

 

b. The images are predictive results for damaged mangosteens at different degrees of hardening. Please provide information if there is a possibility to connect each image with the height from which the real sample was dropped. If yes, please include the information on the height of the drop.

Our response to the respected reviewer: Thank you very much for your kind comment.

No, the height for the drop experiment of each sample wasn’t recorded in this study. The experiment didn’t have a plan for presenting the relation between height and firmness. The variation of firmness of samples was desired for establishing the model that was the main requirement in the drop experiment.

 

6. If possible, please consider widening the analysis of mixed groups for sizes that assure statistical significance of firmness measured with a texture analyzer (for instance a mixed group of sizes 1 and 2 or 1 and 3)

Our response to the respected reviewer: Thank you very much for your kind comment. We agree with your useful comment, but we didn’t have a plan for this case in our experiment. They were only a group of size 3 and a mixed-size group of size 3 and 4 that were used for the drop experiment in order to obtain the variation of firmness for establishing the models and used the results from these two groups for comparing the performance of the models in this study. We agree with your useful comment, but we didn’t have a plan at the first time for this case in our experiment.

 

I believe that considering my comments no. 1-5 can improve the comprehensiveness of the manuscript and introducing the results mentioned in my comment no. 6 would expand the understanding of the proper model selection.

Our response to the respected reviewer: Thank you for your useful comments and suggestions on our manuscript. We have modified the manuscript as your comments and suggestions.

 

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Here is the Reviewer comments:

1.    Interest Relevance
    The article under review lacks significant interest to a broad readership due to its lack of real-time applicability. Consequently, its utility appears limited.

2.    Accuracy and Comparison
    In previous studies, existing applications demonstrate real-time strawberry detection utilizing Hyperspectral Imaging (HSI). Thus, the article requires substantial revisions to ensure accuracy. Failure to address this concern may result in rejection.

3.    Introduction Depth
    The introductory section pertaining to mangosteen is insufficiently detailed, requiring expansion for clarity and comprehensiveness.

4.    Sample Preparation Detail
    The manuscript lacks critical information regarding sample preparation, specifically regarding the calculation methodology for sample allocation across groups. Details such as the number of mangosteens per group and their uniform distribution need clarification.

5.    Materials and Methods Clarity
    The materials and methods section presents notable ambiguities, leading to an overall rejection of the manuscript. It is imperative to rectify these issues. Furthermore, the study's methodology would benefit from a minimum of 500 samples per group with a replication rate of at least three.

Comments on the Quality of English Language

 Minor editing of English language required.

Author Response

Please see the attachment.

Responses to the respected editor and reviewers:

 

We would like to thank all respected reviewers for valuable comments that have helped us to improve our manuscript.  Changes have been highlighted in yellow within the manuscript.  The revised manuscript has been edited the language by the native English speaker again.  We hope you accept these corrections. Our point-by-point answers to the comments are in the following:

 

Reviewer 2

The Reviewer comments:

 

1. 1. Interest Relevance

    The article under review lacks significant interest to a broad readership due to its lack of real-time applicability. Consequently, its utility appears limited.

Our response to the respected reviewer: Thank you very much for your kind comment for bringing up this point. The manuscript has been improved as your suggestion.

 

Line 33-38: Mangosteens are exported worldwide and global exports were reported to be about    2.3 million tonnes in 2023. The export price could fluctuate considerable, but it was among the most expensive fruits on export markets (in 2023 it peaked at $1,870 per tonne) [2] therefore,   

a real-time sorting system for monitoring quality would be a major contributor to the export industry.

 

Reference:

 

[2] FAO. Major Tropical Fruits. Available online: https://www.fao.org/3/cc9308en/cc9308en.pdf. (accessed on 26 February 2024)

 

Line 58-60: Therefore, a nondestructive method, which could be used to detect individual mangosteens with the physiological disorder “hardening” in real-time in an online grading system would be a major asset to the export industry.

 

Line 333-336: Also, the results from this study can be utilized for improving the grading system as well as perhaps being applied for other fruit that can be graded using real-time online inspection for screening fruit during grading and distribution.

 

2. 2. Accuracy and Comparison

    In previous studies, existing applications demonstrate real-time strawberry detection utilizing Hyperspectral Imaging (HSI). Thus, the article requires substantial revisions to ensure accuracy. Failure to address this concern may result in rejection.

Our response to the respected reviewer: Thank you very much for your kind comment for bringing up this point.

There are publications that reported the accuracy of the models for predicting firmness of fruit. The results of published papers are substantial that can be used to compare with the acquired accuracy of our study.

 

Technique	Samples	Predicted variable	Accuracy of the model	References
NIRs	Apple	Firmness	Rp= 0.48	Lu, R. et al. (2000)
NIRs	The European pear	Firmness	Rp2= 0.63	Wang, J. et al. (2017)
Vis–NIR	Apple	Firmness	Rp2= 0.76	Ma, T. et al. (2021)
NIRS	Nanguo pears	Firmness	Rp2= 0.69	Yu, Y.; Yao, M. (2022)

Remarks:  For our manuscript, the accuracy of the models for predicting firmness obtained Rp= 0.87.

 

References:

 

[58] Lu, R.; Guyer, D. E.; Beaudry, R. M. Determination of firmness and sugar content of apples using near-infrared diffuse reflectance. J. Texture Stud. 2000, 31 (6), 615-630. https://doi.org/10.1111/j.1745-4603.2000.tb01024.x

 

[59] Wang, J.; Wang, J.; Chen, Z.; Han, D. Development of multi-cultivar models for predicting the soluble solid content and firmness of European pear (Pyrus communis L.) using portable vis–NIR spectroscopy. Postharvest Biol. Technol. 2017, 129, 143-151. https://doi.org/10.1016/j.postharvbio.2017.03.012

 

[60] Ma, T.; Xia, Y.; Inagaki, T.; Tsuchikawa, S. Rapid and nondestructive evaluation of soluble solids content (SSC) and firmness in apple using Vis–NIR spatially resolved spectroscopy. Postharvest Biol. Techno. 2021, 173, 111417. https://doi.org/10.1016/j.postharvbio.2020.111417

 

[61] Yu, Y.; Yao, M. A portable NIR system for nondestructive assessment of SSC and firmness of Nanguo pears. LWT-Food Sci. Technol. 2022, 167, 113809. https://doi.org/10.1016/j.lwt.2022.113809

 

 

The manuscript has been improved as your suggestion.

Line 295-297: The accuracy of the acquired model for firmness in this study was accepted when compared with previous reports for the models on apples and pears [58-61],

 

3. 3. Introduction Depth

    The introductory section pertaining to mangosteen is insufficiently detailed, requiring expansion for clarity and comprehensiveness.

Our response to the respected reviewer: Thank you very much for your kind suggestion. The introductory section pertaining to mangosteen has been improved as your suggestion.

 

Line 33-38: Mangosteens are exported worldwide and global exports were reported to be about    2.3 million tonnes in 2023. The export price could fluctuate considerable, but it was among the most expensive fruits on export markets (in 2023 it peaked at $1,870 per tonne) [2] therefore,     

a real time sorting system for monitoring quality would be a major contributor to the export industry.

 

Reference:

 

[2] FAO. Major Tropical Fruits. Available online: https://www.fao.org/3/cc9308en/cc9308en.pdf. (accessed on 26 February 2024)

 

Line 44-47: Three major disorders of mangosteen fruits, “hardening”, “translucent flesh” and “gamboge”, cannot be detected by visual inspection and can result in fruit being rejected on export markets or by consumers.

 

Line 58-60: Therefore, a nondestructive method, which could be used to detect individual mangosteens with the physiological disorder “hardening” in real-time in an online grading system would be a major asset to the export industry.

 

4. 4. Sample Preparation Detail

    The manuscript lacks critical information regarding sample preparation, specifically regarding the calculation methodology for sample allocation across groups. Details such as the number of mangosteens per group and their uniform distribution need clarification.

Our response to the respected reviewer: Thank you very much for your kind suggestion. The manuscript regarding sample preparation has been improved as your suggestion.

 

Line 130-133: The number of samples used for the dropping experiment was 560 of which 280 were from the group of size 3 and also 280 from the mixed-size group of size 3 and 4. The height of drop varied from 50 to 100 cm onto a cement floor and was randomly out onto a cement floor.

 

Line 171-177: PLSR and SVMR were used for establishing the calibration models for firmness which were then cross-validated. The samples in the calibration set were used in this analytical process. Statistical data, based on the low root mean square error of cross validation (RMSECV) and high R_cv were used in order to select each optimum calibration model as previously described [46-47]. The accuracy of the calibration models was tested using the samples in the prediction set. The results of accuracy were evaluated by using R_p and RMSEP.

 

Line 198-211: In order to test the possibility that fruit size might affect the accuracy of the models for firmness, a comparison of the models from the different groups of mangosteen    samples that were the models from similar size samples (group A) and different sizes samples (group B) was evaluated. For demonstrating the efficacy and practicality of this method, the number of samples in both groups with the same size (N=280) was used for establishing the models for comparison. For this, two groups of samples were used a similar size group of size 3 (group A) and a mixed-size group of size 3 and 4 (group B). The number of samples in the calibration set for both group A and group B was 196 and for the prediction set for both group A and group B the number was 84. The firmness of the samples from group A varied between 4.51 and 48.55 N and for group B between 3.82 and 49.87 N. The standard deviations of firmness in the calibration set and the prediction set were used considered to ensure the results of model’s performance. The uniform distribution firmness of the samples in the calibration set and the prediction set were quite similar; in group A (11.99 and 12.28 N) and in group B (11.84 and 11.56 N) respectively (Table 2).

 

5. 5. Materials and Methods Clarity

    The materials and methods section presents notable ambiguities, leading to an overall rejection of the manuscript. It is imperative to rectify these issues. Furthermore, the study's methodology would benefit from a minimum of 500 samples per group with a replication rate of at least three.

Our response to the respected reviewer: Thank you very much for your kind suggestion. The manuscript has been improved as your suggestion.

For the comment of the materials and methods section:

 

Line 103-106:  Black and white references were used for NIR-HSI acquisition. The black reference was taken when the lid of the camera was covered and the shutter was closed. The white reference was taken by scanning a Spectral on bar for every measurement.

 

Line 115-120: The mangosteens that were judged to be at the maturity stage IV (reddish-brown), were arranged into their five different size groups (numbered from 1 to 5) based on the standard provided by codex standard for mangosteens (Codex Stan 204-1997) [44]. The firmness of each fruit was determined, and the number of sound mangosteens was 50 samples that were used for considering the among different sizes of mangosteen. The firmness of each fruit was measured using the texture analyzer (TA xt plus),

 

Line 125-133: Dropping was from different heights in order to determine the variation in severity of the hardening disorder on the pericarp of the mangosteens. For this, the texture value of impacted samples could be related to the height of drop based on the severity of the hardening disorder. The distribution of texture values from low to high levels of hardening was required in order to establish the model. The number of samples used for the dropping experiment was 560 of which 280 were from the group of size 3 and also 280 from the mixed-size group of size 3 and 4. The height of drop varied from 50 to 100 cm onto a cement floor and was randomly out onto a cement floor.

 

Line 161-164: The firmness values from the texture measurements could influence the characteristic of acquired spectra, therefore, the firmness value and the spectral data of each sample from the ROI were used as the representative data for establishing the model.

 

Line 171-177: PLSR and SVMR were used for establishing the calibration models for firmness which were then cross-validated. The samples in the calibration set were used in this analytical process. Statistical data, based on the low root mean square error of cross validation (RMSECV) and high R_cv were used in order to select each optimum calibration model as previously described [46-47]. The accuracy of the calibration models was tested using the samples in the prediction set. The results of accuracy were evaluated by using R_p and RMSEP.

 

Line 185-186: Samples of both sound and hardened mangosteens were used for creating the predictive images based on firmness for comparison.

 

 

For the comment of a minimum of samples per group:

There are publications that reported the number of samples for establishing the model.

Technique	Samples	Predicted variable	No. of all samples	No. of samples for calibration	No. of samples for prediction	References
Vis-NIR-HSI	Fennel heads	SSC, DPPH and phenols	136	105	31	Amodio, M. L. et al. (2017)
Vis-NIR-HSI	Fresh-cut potato slices	E. coli	128	91	37	Li, D. et al. (2021)
NIRs	pears	SCC and Firmness	90	60	30	Yu, Y.; Yao, M. (2022).
NIR-HSI	sugarcane stalk	commercial cane sugar (CCS) content	210	140	70	Chiatrakul, J. et al. (2022)
NIR-HSI	tomatoes	Firmness	148	118	30	Zhao M. et al. (2023)
NIR-HSI	Chicken	TVB-N and TVC	240	168	72	Li, X. et al. (2023)
NIR-HSI	Cheese	Moisture Fat Protein Free oil	66 70 66 67	46 49 45 47	20 21 21 20	Karaziack, C. B. el al (2024)
NIRs	Lime	TA	131	98	33	Li, P. el al (2024)

The methodology of our study was similar to the other publications by dividing samples to the calibration and the prediction set. The number of samples in our experiment was reasonably enough when compared with other publications. [There were 280 samples in each group of our study and they were also divided to the calibration set (N= 196) and the prediction set (N = 84)].

 

References:

Amodio, M. L.; Capotorto, I.; Chaudhry, M. M. A.; Colelli, G. The use of hyperspectral imaging to predict the distribution of internal constituents and to classify edible fennel heads based on the harvest time. Comput. Electron. Agric. 2017, 134, 1-10. DOI:10.1016/j.compag.2017.01.005

 

Li, D.; Zhang, F.; Yu, J.; Chen, X.; Liu, B.; Meng, X. A rapid and non-destructive detection of Escherichia coli on the surface of fresh-cut potato slices and application using hyperspectral imaging. Postharvest Biol. Technol. 2021,171, 111352. https://doi.org/10.1016/j.postharvbio.2020.111352

 

Yu, Y.; Yao, M. A portable NIR system for nondestructive assessment of SSC and firmness of Nanguo pears. LWT-Food Sci. Techno. 2022, 167, 113809. https://doi.org/10.1016/j.lwt.2022.113809

 

Chiatrakul, J.; Terdwongworakul, A.; Phuangsombut, K.; Phuangsombut, A. Improved evaluation of commercial cane sugar content in sugarcane stalk using near infrared hyperspectral imaging and stalk axis rotation technique. Biosyst Eng. 2022, 223, 161-173. https://doi.org/10.1016/j.biosystemseng.2022.08.019

 

Zhao, M.; Cang, H.; Chen, H.; Zhang, C.; Yan, T.; Zhang, Y.; Gao, P.; Xu, W. Determination of quality and maturity of processing tomatoes using near-infrared hyperspectral imaging with interpretable machine learning methods. LWT-Food Sci. Technol. 2023, 183, 114861. https://doi.org/10.1016/j.lwt.2023.114861

 

Li, X.; Cai, M.; Li, M.; Wei, X.; Liu, Z.; Wang, J.; Jia, K.; Han, Y. Combining Vis-NIR and NIR hyperspectral imaging techniques with a data fusion strategy for the rapid qualitative evaluation of multiple qualities in chicken. Food Control. 2023,145, 109416.  https://doi.org/10.1016/j.foodcont.2022.109416

 

Karaziack, C. B.; Vidal, C.; Pasquini, C.; Barbin, D. F.; Viotto, W. H. Application of near-infrared hyperspectral imaging for determination of cheese chemical composition. J. Food Compos. Anal. 2024,127, 105994. https://doi.org/10.1016/j.jfca.2024.105994

 

Li, P.; Dong, Y.; Jiang, L.; Du, G.; Shan, Y. Nondestructive prediction of lime acidity with a single scan using two types of near infrared spectrometers and ensemble learning strategy. J Food Eng. 2024, 368, 111917. https://doi.org/10.1016/j.jfoodeng.2023.111917

 

 

Comments on the Quality of English Language

 Minor editing of English language required.

Thank you very much for your kind suggestion. The revised manuscript has been edited the language by the native English speaker again before submission.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I think this research is valuable as applied research for using near-infrared and hyper spectral imaging. I would like to have few comments that I noticed.

 

Line84-85

What is the standard for sizing by weight? Is there a relationship between size based on weight and ripeness?

2.2NIR-HIS acquisition

What reference did you use?

2.3. Firmness measurement

Firmness measurement is difficult to understand. Do you use this measurement method (dropping measurement) because you want to measure non-destructively? Why not perform invasive firmness testing?

Table1 and 2

Even though Group 4 is harder than Group 3, why is Group B having a lower mean firmness?

Line190-191

Group B has less light absorption in the water region, so does it have less water? Is it correct to assume that the fruit is hard because there is less water?

Line204

Why does the required number of latent variables change when the sample condition changes (group A and B)even if the same statistical processing is performed?

Line284-286

What do you think is the reason why PLSR is more accurate than SVR?

 

The results of this study showed that the prediction accuracy was higher when the sample size was uniform. I believe that this measurement was performed excluding large and small samples, but would it be a good method to create a calibration curve for each size when applied in the field such as a fruit sorting plant?

Author Response

Please see the attachment.

Responses to the respected editor and reviewers:

 

We would like to thank all respected reviewers for valuable comments that have helped us to improve our manuscript.  Changes have been highlighted in yellow within the manuscript.  The revised manuscript has been edited the language by the native English speaker again.  We hope you accept these corrections. Our point-by-point answers to the comments are in the following:

Reviewer 3

Comments and Suggestions for Authors

 

I think this research is valuable as applied research for using near-infrared and hyper spectral imaging. I would like to have few comments that I noticed.

Our response to the respected reviewer: Thank you very much for your positive feedback on our manuscript. The manuscript has been improved as your kind comments.

 

1.Line84-85

What is the standard for sizing by weight? Is there a relationship between size based on weight and ripeness?

Our response to the respected reviewer: Thank you very much for your question. There is a reference for a relationship between sizes based on weight but there is no reference for a relationship between size and ripeness.

The standard size of mangosteen based on weight by codex standard for mangosteens (Codex Stan 204-1997) is shown as following:

 

Table Size is determined by the weight

Size code	Weight (in grams)
1	>125
2	100-125
3	76-100
4	51-75
5	30-50

 

The manuscript has been improved.

Line 115-120: The mangosteens that were judged to be at the maturity stage IV (reddish-brown), were arranged into their five different size groups (numbered from 1 to 5) based on the standard provided by codex standard for mangosteens (Codex Stan 204-1997) [44]. The firmness of each fruit was determined, and the number of sound mangosteens was 50 samples that were used for considering the among different sizes of mangosteen. The firmness of each fruit was measured using the texture analyzer (TA xt plus),

 

Reference:

[44] FAO. Standard for mangosteen. (Codex Stan 204-1997). 2005. pp. 2.

 

 

 

 

2. 2. NIR-HSI acquisition

What reference did you use?

Our response to the respected reviewer: Thank you very much for your question. The manuscript has been improved as your suggestion.

 

Line 103-106:  Black and white references were used for NIR-HSI acquisition. The black reference was taken when the lid of the camera was covered and the shutter was closed. The white reference was taken by scanning a Spectral on bar for every measurement.

 

3. 3. Firmness measurement

Firmness measurement is difficult to understand. Do you use this measurement method (dropping measurement) because you want to measure non-destructively? Why not perform invasive firmness testing?

Our response to the respected reviewer: Thank you very much for your questions and your kind attention. The drop experiment was designed for creating hardening mangosteens. The degree of hardening was related to the severity of hardening disorder based on the height of drop. The actual firmness value of each mangosteen was measured by using the texture analyzer (TA xt plus) that was an invasive measurement. The data of firmness and spectral data from HSI were used for establishing model for predicting firmness value. Therefore, scanning a mangosteen by HSI and used the acquired model that was a non-destructive technique for predicting the firmness.

 

Line 120: The firmness of each fruit was measured using the texture analyzer (TA xt plus),

 

Line 125-133: Dropping was from different heights in order to determine the variation in severity of the hardening disorder on the pericarp of the mangosteens. For this, the texture value of impacted samples could be related to the height of drop based on the severity of the hardening disorder. The distribution of texture values from low to high levels of hardening was required in order to establish the model. The number of samples used for the dropping experiment was 560 of which 280 were from the group of size 3 and also 280 from the mixed-size group of size 3 and 4. The height of drop varied from 50 to 100 cm onto a cement floor and was randomly out onto a cement floor.

 

Line 161-164: The firmness values from the texture measurements could influence the characteristic of acquired spectra, therefore, the firmness value and the spectral data of each sample from the ROI were used as the representative data for establishing the model.

 

4. 4. Table1 and 2

Even though Group 4 is harder than Group 3, why is Group B having a lower mean firmness?

Our response to the respected reviewer: Thank you very much for your questions and your kind attention. Table 1 presented the firmness of sound mangosteens while Table 2 presented the firmness of hardening mangosteens. The hardening mangosteens were randomly created by the drop experiment. Therefore, the firmness of the hardening mangosteens in Table 2 was varied and wasn’t related to those of samples in Table 1.

 

 

 

5. 5. Line190-191

Group B has less light absorption in the water region, so does it have less water? Is it correct to assume that the fruit is hard because there is less water?

Our response to the respected reviewer: Thank you very much for your questions and bringing up this point. In Figure 4, the average absorbance spectrum of samples in Group B was less than Group A’s in a whole wavelength, not only in the water region. The absorbance in the whole wavelength of each sample might be different because of many factors such as the difference of surrounding, features of samples etc. Therefore, the comparison of quantity of water or other components of mangosteen in each group couldn’t be explained by acquired original spectra.

 

6. 6. Line204

Why does the required number of latent variables change when the sample condition changes (group A and B) even if the same statistical processing is performed?

Our response to the respected reviewer: Thank you very much for your questions and bringing up this point. The different number of latent variables of each group could be possible because the characteristics of samples in each group were different. Also, the suitable number of latent variables was selected in order to avoid under or over fitting of the model. The manuscript has been improved as your kind comments.

 

Line 233-235: In order to avoid underfitting or overfitting of the model, the graph of RMSECV versus the latent variables was tested. For this the latent variables at the point of lowest RMSECV were selected.

 

7. 7. Line284-286

What do you think is the reason why PLSR is more accurate than SVR?

Our response to the respected reviewer: Thank you very much for your questions and bringing up this point. Both PLSR and SVR are methods for use to establish the model for predicting the predicted variable from several independent variables. PLSR is applied to find the linear model from the relation between 2 metrics (X and Y) that are projected to new spaces. SVR is applied to build the model based on calculating a distance metric among data vectors that the linear model is constructed in the hyperplanes. I think the accuracy of acquired models from PLSR and SVR is varied based on the characteristics of samples and the factors that affect the relation between predicted variable and independent variables.

 

8. 8. The results of this study showed that the prediction accuracy was higher when the sample size was uniform. I believe that this measurement was performed excluding large and small samples, but would it be a good method to create a calibration curve for each size when applied in the field such as a fruit sorting plant?

Our response to the respected reviewer: Thank you very much for your kind comment.

Yes, the results showed that the performance of sorting units would be better when using for each size. For application, mangosteens should be graded for each size first and then mangosteens will pass though the sorting units at the next step. The results from this study can be applied for improving the sorting system of the fruit factory that can be utilized for screening the high quality of fruit before sending to consumers. The manuscript has been improved as your kind comments.

 

Line 333-336: Also, the results from this study can be utilized for improving the grading system as well as perhaps being applied for other fruit that can be graded using real-time online inspection for screening fruit during grading and distribution.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The Authors need to revise the manuscript for speed onto the plagiarism detection with iThenticate.

I recommend aiming for an acceptance rate of approximately 25%.

Once the revisions are complete, I will check your manuscript again.

 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

 Minor editing of English language required.

Author Response

Responses to the respected reviewer (Round 2):

We appreciate your valuable comment very much. The manuscript has been improved. We hope these changes will meet with your approval. The new changes are highlighted in red front.

 

Reviewer 2

Comments and Suggestions for Authors

The Authors need to revise the manuscript for speed onto the plagiarism detection with iThenticate. I recommend aiming for an acceptance rate of approximately 25%. Once the revisions are complete, I will check your manuscript again.

Our response to the respected reviewer: Thank you very much for your comment on our manuscript. The revised manuscript has been improved.

 

Line 21-23: Using partial least squares regression (PLSR) the correlation coefficient of prediction (R_p) was 0.87 and the root mean square error of prediction (RMSEP) was 6.25 N.

 

Line 24-26: From the data it was concluded that NIR-HSI can potentially be used to visualize hardening of individual mangosteens based on their predictive images.

 

Line 31-33: They are grown commercially in many tropical areas including in Thailand, Indonesia, Vietnam, Malaysia and Philippines, with global exports of about 2.3 million tonnes in 2023 [2].  

 

Line 40-42: For optimum quality, mangosteens are harvested at stage III and usually eaten when they have ripened to stage IV.

 

Line 60-63: A possible method to achieve this is by using hyperspectral imaging (HSI). HSI is a technique that is rapid, non-destructive and combines conventional spectroscopy and imaging techniques in order to acquire both spectral and spatial information in order to predict certain qualities of products [12].

 

Line 75-77: measuring the quality in longans [35], determining of the TVB-N and TVC in chicken [36], identifying adulterated chickpea flour [37], assessing nitrite content in Vienna sausages [38] and detecting the maturity of pineapples [39].

 

Line 77-79: Near infrared hyperspectral imaging (NIR-HSI) has previously been used for non-destructive measurement of textural characteristics of apples [40],

 

Line 114: The firmness of each fruit was determined using a texture analyzer (TA xt plus),

 

Line 116-118: The TA xt plus was fitted with a 2 mm diameter cylindrical plunger and applied at a speed of 10 mm/s to a depth of 5 mm into the skin of each fruit.

 

Line 129-131: Each fruit was dropped in such a way that its calyx was horizontally orientated in order that the area of impact was consistent. The floor was covered with talcum powder

 

Line 136-137: Directly after scanning, the firmness of each fruit was measured at the impacted area using the texture analyzer.

 

Line 139-141: In order to determine the relationship between firmness of sound mangosteens and their size, the data from the firmness of different sizes (1 to 5) of the same maturity stage (reddish-brown) were statistically analyzed using the Duncan's Multiple Range Test.

 

Line 181-182: Predictive images of samples of both sound and hardened mangosteens were used for comparison based on firmness.

 

Line 183-186: For statistical analysis, the ChemaDAQ software (Specim, Spectral Imaging Ltd, Oulu, Finland), the UmBio Evince hyperspectral image analysis software (Prediktera Evince, version 2.7.5, Sweden) and the Unscrambler X (CAMO, version 10.5.1, Norway) were used.

 

Line 211-212: Firmness values are given as means ± standard deviation. Where values are followed by the same letter, they were not significantly different (p ≤ 0.05) using the Duncan's Multiple Range Test.

 

Line 216-219: The average spectral data in the region of 935-1720 nm from the ROI of samples from both groups were performed in order to consider the spectrum of the mangosteens (Figure 4). The results showed that spectra of both groups contained the main absorbance peaks at around 970, 1200 and 1450 nm. Since these are the absorption peaks of water [56],

 

Line 224-226: In order to achieve the optimum calibration model spectral pretreatments for    firmness using PLSR (Table 3) and the optimum calibration model for firmness by SVMR (Table 4) were investigated. 

 

Line 261-262: The best results for both group A and group B were from the Savitzky-Golay first derivative differentiation spectral pretreatment.

 

Line 289-290: The PLSR model from group A was used to create the predictive images since it was shown to provide the best model.

 

Line 302: The predictive images clearly showed different colors within the image,  

 

Line 315-317: The results indicated that a partial least squares regression model for non-destructively could accurately be used to predict firmness of individual mangosteens and gave more accurate results compared to a support vector machine regression.

 

Line 325-327: It has the potential for incorporation within a commercial non-destructive grading system for detecting the degree of hardening of mangosteens, which is important for modern automated fruit grading systems.

 

The plagiarism of the main content of the revised manuscript has been rechecked by Turnitin and the result of a similarity index was 21%

Thank you very much for your kind suggestion. The English language of the revised manuscript has been edited again by the native English speaker before submission.

Please see the attachment

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript was revised carefully as the Reviewer expected. I accept the manuscript as the present form.