Next Article in Journal
Spatial Autocorrelation Analysis of CO and NO2 Related to Forest Fire Dynamics
Previous Article in Journal
Evaluation of Pedestrian-Perceived Comfort on Urban Streets Using Multi-Source Data: A Case Study in Nanjing, China
 
 
Article
Peer-Review Record

VE-GCN: A Geography-Aware Approach for Polyline Simplification in Cartographic Generalization

ISPRS Int. J. Geo-Inf. 2025, 14(2), 64; https://doi.org/10.3390/ijgi14020064
by Siqiong Chen 1,2, Anna Hu 3, Yongyang Xu 2,4,*, Haitao Wang 1 and Zhong Xie 4
Reviewer 1:
Reviewer 2:
Reviewer 3: Anonymous
ISPRS Int. J. Geo-Inf. 2025, 14(2), 64; https://doi.org/10.3390/ijgi14020064
Submission received: 23 December 2024 / Revised: 3 February 2025 / Accepted: 4 February 2025 / Published: 6 February 2025

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Polyline simplification is a critical process in cartographic generalization. This research designed a novel approach—a joint vertex-edge features graph convolution network (VE-GCN) for polyline simplification. The research is innovative and a lot of experiments have been done to prove the effectiveness and advancement of the proposed method. However, there are still some problems should be explained clearer.

1. The research motivation should be elaborated in detail in the abstract.

2. The specific manifestation of the proposed method should be clearly stated in the abstract.

3. Although the introduction addresses the current methods' shortcomings, it is suggested to further summarize and condense this discussion.

4. In the "Related Works" section, the research status of the line simplification methods in cartographic generalization is not comprehensive enough. A more thorough analysis is recommended, such as elaborating on the limitations of traditional geometric simplification and deep learning methods in geographic feature extraction.

5. In the methodology section, the differences in edge and point features are not described in enough detail. It is suggested to provide a clearer explanation of the implementation details.

6. There are several minor issues in the paper, such as inconsistent use of technical terms (e.g., "convolution" vs. "convolutional"), incomplete figure captions (e.g., the sub-figures in Fig. 3 lack labels like (a), (b)), and some blurry images.

7. The font size in the figures is inconsistent, and certain language descriptions are unclear. It is recommended to thoroughly check the entire manuscript and make necessary revisions.

Comments on the Quality of English Language

Some language descriptions are unclear, it is recommended to thoroughly check the entire manuscript and make necessary revisions.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Simplification of polylines is one of the most extensively studied areas in the algorithmization of cartographic generalization. There are several dozen algorithms (and if we include various modifications, more than 100) that perform this process of simplification. The presented article proposes a relatively innovative method for simplifying the course of polylines and compares this method with selected algorithms for simplification. A description and demonstration of the proposed method is generally well-described and undoubtedly deserves further discussion in the generalization community . However, I have several comments regarding the subject of the article:

1. The introductory characterization of simplification algorithms is somewhat oversimplified. While it is true that polylines consist of edges and vertices, and edge complexes can be perceived as bends, it must be noted that the problem of defining and identifying bends is a relatively complex matter that cannot be encompassed under a single mechanism. Moreover, there are algorithms that directly deal with edges, so the inclusion of edges for simplifying the course of polylines is not such an extremely innovative mechanism.

2. It is not evident from the article whether the input data underwent any homogenization, which is crucial because the proposed mechanism depends on the size of the edges, which is often influenced by oversampling.

3. The degree of similarity expressed through various distance measurements between the original and simplified shapes may not be entirely indicative of the effectiveness of simplification for maps at different scales. Whether using Euclidean, Hausdorff, or Fréchet distance, the results may not fully represent the practical impact of the simplification.

4. Although the article compares several scenarios of polyline drawings, it should be noted that the problem of polyline simplification lies in the fact that many methods perform very well on certain types of polylines but not as well on others. It would therefore be useful to see many more examples of significantly different polyline courses.

5. What I find most lacking in assessing the results of the method is some estimate of the computational complexity of processing the dataset. This is again a key factor when selecting algorithms for polyline simplification, as, as mentioned, there is a wide range of these algorithms with varying complexity and efficiency.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The authors of the paper introduce, in the context of cartographic generalization, a new methodology (named VE-GCN) for the simplification of polylines (digital representations of cartographic lines) based on a graph convolutional network with joint vertex-edge features. Authors argue that their approach retains the shape of polylines at the transition from large to small scales. 

 

I appreciate that the authors clearly articulate all aspects of their research. They explain that their motivations are driven by the goal of overcoming specific limitations of existing line simplification algorithms. The literature review they provide on related works is sufficient. They thoroughly describe the proposed methodology and compare it against the line simplification algorithm of Douglas and Peucker. The results are well discussed and summarized in the conclusions. The tables and figures effectively aid readers in understanding the proposed methodology.

 

Additionally, I suggest the authors to consider the following points:

1.   The phrase in lines: 36-37 is recommended for removal, as map generalization has been crucial to all maps, from the earliest created to the present.

2.   After the phrase in lines: 139-141 is recommended to authors to add a short comment for the line simplification algorithm known as “bend simplification” algorithm:

Wang, Z., and Müller, J.-C., 1998. “Line Generalization Based on Analysis of Shape Characteristics”, Cartography and Geographic Information Systems, 25(1): 3-15.

3.   At the end of sub-section: “2.2 Machine learning-based simplification methods” (lines: 196-211) should be removed.

4.   It is recommented to authors to clarify how “trigonometric functions” (in the phrase located in lines: 257-258) can be treated as attribute or spatial characteristic of the edges along a polyline.

5.   The definitions of the variables in relation (7), at the end of section: “3. Materials and Methods” are missing.

6.   Authors, in order to evaluate the results, they define and apply a metric named: “differential distance” (in line: 442). This specific metric has been widely used by the cartographic community with the term: areal displacement. See for example:

White, E.R., 1985. “Assessment of Line-Generalization Algorithms Using Characteristic Points”, Cartography and Geographic Information Science (Former: The American Cartographer), 12(1): 17-28.

McMaster, R.B., 1986. “A Statistical Analysis of Mathematical Measures for Linear Simplification, Cartography and Geographic Information Science (Former: The American Cartographer), 13(2): 103-116.

Additionally, authors may consider several limitations of “areal displacement” metric introduced in:

Kronenfeld, B.J., and Deng, J., 2019. “Between the Lines: Measuring Areal Displacement in Line Simplification”, Advances in Cartography and GIScience of the ICA, vol. 1. (https://doi.org/10.5194/ICA-ADV-1-9-2019).

7.   It is recommended to authors to clarify at the end of sub-section: “4.2 Road simplification experiment” (line: 459), if they are refering to Douglas and Peucker line simplification algorithm.

8.   It is recommended to authors to discuss in more detail on how a user can tune the parametrers of VE-GCN line simplification approach in order to achive a good result at a specific change of scale.

9.   Authors may consider the following issues related to the section: "References":

-  It seems that there something wrong with the title of the Reference 10: “I am using Genetic Algorithms for Solving Problems in Automated Line Simplification”, it should be checked. Additionally the name of the the journal should be corrected to: “Sinica”.

-  In Reference 19: The issue number should be added. So, “30” in line: 776 should be replaced by “30(1)”.

-  In Reference 20: The title of the paper “Algorithms for Objective Generalization of Line Features Based on the Natural Principle” in line: 777 should be replaced by “Algorithms for the Automated Line Generalization Based on Natural Principle of Objective Generalization”.

-  Reference 21: should be replaced with a more extensive publication of the same authors in the same topic:

Nakos B., Mitropoulos V., 2005. “Critical Points Detection Using the Length Ratio (LR) for Line Generalization”, Cartographica, 40(3): 35-51.

-  In Reference 22: The year of publication should be corrected. So, “2000” in line: 782 should be replaced by “1989”.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop