Large-Scale and High-Accuracy Phenotyping of Populus simonii Leaves Using the Colony Counter and OpenCV
Round 1
Reviewer 1 Report
This manuscript assess the accuracy of image-based method to estimate area and length of individual Populus simonii leaves using the colony counter Scan 1200 and OpenCV library. The proposed method includes image pre-processing, image segmentation, and object contour detection. The paper is in a good shape to be published, the subject is quite interesting, I enjoyed reading the manuscript. I have only minor comments that the authors should consider in the revised version of the paper.
1. Please, make the objective of the study clearer. In the Abstract, the objective was stated as: to presented an automatic image-based workflow to obtain accurate estimations for basic leaf characteristics from a hundred of Populus simonii pictures. In the Introduction section, the objective was stated as: to estimate the leaf morphological traits from the P. simonii images using a three-step workflow, composed of image pre-processing, image segmentation, and contour-based prediction. If I am not wrong, the main objective of the study was to evaluate the performance of different image segmentation techniques in estimating morphological parameters.
2. In the Abstract, please, provide the morphological parameters that were considered by the authors, that is, area, length, width, and ratio length/width.
3. In the Introduction section, the authors stated that they considered four image segmentation strategies (threshold-based, region-based, edge-based, and K-Means-based algorithms) (L67-68). However, in L99-102, the authors stated seven algorithms: Chan-Vese segmentation, OSTU thresholding, iterative thresholding, adaptive thresholding, Canny edge detection, K-means clustering, and Watershed algorithm. In the Abstract, Discussion and Conclusion sections, the authors reported that considered six approaches. Please, be consistent in the text regarding this matter.
4. Why the leaf that presented anomalously high standard deviation of 4.90 was not excluded or replaced?
5. In Figure 1, the inclusion of length/width ratio in the accuracy assessment presents secondary importance since the authors are already evaluating both length and width. I missed contour data in this figure.
6. In Table 1, please, provide the unit of measurement (second).
7. In Figure 8, there is a noticeable trend of underestimation of length measurements that should be reported in the text.
8. The pixel size of LI-3100 C was informed in the text (0.1 mm x 0.1 mm). What was the resolution of images captured by the Scan 1200?
9. The authors forgot to provide the most important result of the paper in the Abstract and Conclusion sections: ultimately, which segmentation algorithm is recommended?
There is room to improve the English writting, but it is unnecessary asking to go for a professional service. The staff from editorial board can improve it easily.
Author Response
Dear reviewers,
Thanks very much for your constructive comments on our manuscript. These comments are all valuable and very helpful for revising and improving our paper, as well as providing important guidance for our research. We have carefully paid attention to comments and have made revisions, which we hope meet with approval. The responses to your comments are as follows.
Response to Reviewer #1
Comment (1). Please, make the objective of the study clearer. In the Abstract, the objective was stated as: to presented an automatic image-based workflow to obtain accurate estimations for basic leaf characteristics from a hundred of Populus simonii pictures. In the Introduction section, the objective was stated as: to estimate the leaf morphological traits from the P. simonii images using a three-step workflow, composed of image pre-processing, image segmentation, and contour-based prediction. If I am not wrong, the main objective of the study was to evaluate the performance of different image segmentation techniques in estimating morphological parameters.
Revision (1): We have made some revisions in the Abstract section.
Comment (2). In the Abstract, please, provide the morphological parameters that were considered by the authors, that is, area, length, width, and ratio length/width.
Revision (2): We have added these parameters in the Abstract section.
Comment (3): In the Introduction section, the authors stated that they considered four image segmentation strategies (threshold-based, region-based, edge-based, and K-Means-based algorithms) (L67-68). However, in L99-102, the authors stated seven algorithms: Chan-Vese segmentation, OSTU thresholding, iterative thresholding, adaptive thresholding, Canny edge detection, K-means clustering, and Watershed algorithm. In the Abstract, Discussion and Conclusion sections, the authors reported that considered six approaches. Please, be consistent in the text regarding this matter.
Revision (3): We have made revisions in the Abstract, Result, Discussion, and Conclusion section.
Comment (4): Why the leaf that presented anomalously high standard deviation of 4.90 was not excluded or replaced?
Revision (4): A very small number of outliers (~1%) with high standard deviation are inevitably appear in large-scale area measurement using LI-3100C owing to its track-type conveyor and leaf’s non-flat surface. We want to display the reality property of leaf area values measured using LI-3100C.
Comment (5): In Figure 1, the inclusion of length/width ratio in the accuracy assessment presents secondary importance since the authors are already evaluating both length and width. I missed contour data in this figure.
Revision (5): In Figure 1, four morphological traits (e.g., ratio of length/width, length, width, & area) of P. simonii leaves were measured using LI-3100C and plastic clear ruler. Nevertheless, we did not determine the contour for P. simonii leaves.
Comment (6): In Table 1, please, provide the unit of measurement (second).
Revision (6): We have added the unit of running time (seconds) in Table 1.
Comment (7): In Figure 8, there is a noticeable trend of underestimation of length measurements that should be reported in the text.
Revision (7): We have added in corresponding text.
Comment (8): The pixel size of LI-3100 C was informed in the text (0.1 mm x 0.1 mm). What was the resolution of images captured by the Scan 1200?
Revision (8): The resolution of images captured by the Scan 1200 (0.097 mm x 0.097 mm) was in close proximity to that of LI-3100 C.
Comment (9): The authors forgot to provide the most important result of the paper in the Abstract and Conclusion sections: ultimately, which segmentation algorithm is recommended?
Revision (9): We have added in the Abstract and Conclusion section.
Reviewer 2 Report
Dear authors,
please find attached file my comments
Best
Comments for author File: 
Comments.pdf
Author Response
Response to Reviewer #2
Dear reviewers,
Thanks very much for your constructive comments on our manuscript. These comments are all valuable and very helpful for revising and improving our paper, as well as providing important guidance for our research. We have carefully paid attention to comments and have made revisions, which we hope meet with approval. The responses to your comments are as follows.
Comment (1): Discussion: The discussion section should go beyond merely reiterating the results and include a broader interpretation of the findings. The authors should discuss how the use of the Colony counter and OpenCV for phenotyping can impact the field of plant sciences. Consider addressing questions such as: - How does the proposed methodology compare to traditional phenotyping methods in terms of accuracy, efficiency, and scalability?
- What are the potential applications of this approach beyond Populus simonii leaves? Could it be adapted for other plant species or different morphological traits?
- How does this research contribute to advancing the understanding of plant development, growth, and response to environmental factors?
Revision (1): We have made some revisions in the Discussion section.
Comment (2): Statistics Information: Regarding the statement about the repeatability and unbiased measure for total foliar area using the LI-3100C (lines 111-113), the authors should provide the specific statistics information (e.g., mean, standard deviation, units) to support their claim. This will allow readers to assess the reliability of the measurements more accurately.
Revision (2): The statistics information should be added in Supplementary Table S1.
Comment (3): Correct Sentence Structure: The sentence in line 125, "The area, length, width, and length/width ratio for all the leaves ranged in size from 9.95 to 33.76 cm2, from 4.2 to 7.7 cm, from 2.9 to 6.9 cm, and from 0.89 to 2.41, respectively (Figure 1)," could be rephrased for better clarity. Consider separating the ranges for each parameter to avoid confusion.
Revision (3): The sentence has been rewritten.
Comment (4): Description of Genotypes: The authors mentioned a relatively high level of intra-individual or intra-genotype variability in leaf traits (line 117), but the genotypes are not adequately described in the "materials and methods" section. Providing more details about the genotypes used in the study will help readers better understand the sources of variability in the leaf traits.
Revision (4): The additional information for four genotypes have been added in the "materials and methods" section.
Comment (5): Figure 1: It is essential to clarify which plant belongs to which genotype in Figure 1. Providing clear labels or a legend will improve the figure's interpretability.
Revision (5): We have made some minor revisions to the legend in Figure 1.
Comment (6): Length to Width Ratio: The introduction of the length to width ratio in the Figure 1 caption creates difficulty in understanding. It would be beneficial to include a clear description of this ratio and its significance in the "materials and methods" section instead.
- Are there any limitations or challenges associated with the proposed method, and how can they be addressed in future research?
Revision (6): We have given a supplementary description of the length to width ratio in the "materials and methods" section. Additionally, the limitations of the image-based phenotype workflow (e.g., hulking body, AC (alternating current) power demand, and small view-field of colony counter Scan1200) and its corresponding solution have already been mentioned in the Discussion section.
Comment (7): Inclusion of Images: To aid readers in better understanding the sampling and photography process, the authors could consider including relevant images in the manuscript. Visual representation of the experimental setup, sample collection, and image processing steps will enhance the readers' comprehension of the methodology.
Revision (7): Figure 2 has been redrawn to better explain the image-based phenotyping procedure to the reader. In addition, the description of the figure has been modified.
Author Response File: 
Author Response.docx
Round 2
Reviewer 2 Report
Dear Authors,
thank you for addressing all my comments
Sincerely,
