Improvement of Region-Merging Image Segmentation Accuracy Using Multiple Merging Criteria

Round 1

Reviewer 1 Report

General comments

The paper is focused on developing image segmentation framework to improve the art of segmentation and the assessment of segmentation accuracy. It used QuickBird satellite images to conduct the experiment. Generally, the paper is well written in terms of structure and technical details. That said, the discussion section needs to be improved and be situated within the context of existing studies. The authors have not presented a case on the significance of the results on the subject of OBIA.

The first paragraph of the discussion reads more as an explanation of methods than an explanation of results. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This manuscript addresses image segmentation through optimizing region-merging based on multiple merging criterions. An iteration optimization approach is proposed. I this this manuscript should be substantially revised before it is considered for publication.
[1] Image segmentation usually extracts a region of interest or divides the scene into meaningful regions. However, the experimental results in this manuscript do not provide these. Only hundreds of sectors that are meaningless are provided. 
[2] Even in areas of solid color objects with plain textures, such as building roofs, the boundaries of segmented areas are not clear. The segmentation results may not be useful for post-processing such as object detection and recognition.
[3] The organization of abstracts and conclusions is unusual. The methodology is usually presented in the abstract and the conclusion usually includes a summary, highlights and future plans.
[4] The author claims that optimization is used for regional merging. What are the criteria for optimization? 
[5] It is difficult to evaluate the overall performance of the segmentation result with the proposed method. Although the authors give some examples, they look very trivial compared to the whole scene.
[6] Only some illustrations and pseudocodes are provided with lenghy explanation. Authors should come up with more suitable equations.  

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

1. An analysis of time complexity is missing.
2. Comparisons with state-of-the-art methods are missing. For example the proposed method can be compared with [1-2].

[1] Su, T., Liu, T., Zhang, S., Qu, Z., & Li, R. (2020). Machine learning-assisted region merging for remote sensing image segmentation. ISPRS Journal of Photogrammetry and Remote Sensing, 168, 89-123..

[2] Su, Tengfei. "Scale-variable region-merging for high resolution remote sensing image segmentation." ISPRS Journal of Photogrammetry and Remote Sensing 147 (2019): 319-334.

3. If it is possible, please apply your method on more public datasets to get more results and you may add more comparisons with state-of-the-art methods.
4. Please, provide a webpage (in the paper) that will include a link to the used dataset and your results for people that are interesting for comparisons.
5. You may provide some negative examples to explain the cases where the proposed method fails. This will help the reader to understand also the cases where the proposed method give high performance results – that can be also provided and analyzed.
6. In addition, in order to improve your related work, you can cite the following related works.

[1] Su, Tengfei. "Scale-variable region-merging for high resolution remote sensing image segmentation." ISPRS Journal of Photogrammetry and Remote Sensing 147 (2019): 319-334.

[2] I. Grinias , C. Panagiotakis and G. Tziritas, MRF-based Segmentation and Unsupervised Classification for Building and Road Detection in Peri-urban Areas of High-resolution, ISPRS Journal of Photogrammetry and Remote Sensing, vol. 122, pp. 145-166, 2016.

[3] A.M. Braga, R.C.P. Marques, F.A.A. Rodrigues, F.N.S. Medeiros,  A median regularized level set for hierarchical segmentation of SAR images, IEEE Geosci. Remote Sens. Lett., 14 (2017), pp. 1171-1175

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Authors revised the manusript according to the reviewer's comments. I think it can be published in current form.