Slice-Based Instance and Semantic Segmentation for Low-Channel Roadside LiDAR Data
Round 1
Reviewer 1 Report
The topic addressed is relevant and has been the subject of research in recent years. In general, the article is well written. However, two points should be improved: 1) The segmentation process uses several thresholds. In this sense, it is necessary to describe how the values of these thresholds were selected. This helps the reader to have an idea of the value to be adopted, if he applies the segmentation method using a different data set or study area; 2) In order to verify the robustness of the proposed method, more experiments should be considered. As a suggestion, I recommend running the proposed method in another study area.Author Response
Dear Reviewer:
We wish to express our very deep appreciation, and the appreciation of all of us, to your great efforts and suggestions for our manuscript. They are valuable and very helpful for revising and improving our paper, as well as the important guiding to our researches.
The attachment is a point-to-point response to your comments and the responses are in red. The modification marked in red in revised version.
We tried our best to improve the manuscript and made some changes in the manuscript. These changes will not influence the content and framework of the paper. And here we did not list the changes but marked in red in revised paper. We appreciate for your warm work earnestly, and hope that the correction will meet with approval. Thank you for your time and patience. I look forward to receiving your letter.
Once again, we would like to thank you for the constructive comments and suggestions. Please feel free to contact us with any questions. We are looking forward to your reply.
Yours sincerely,
Authors
Author Response File: 
Author Response.docx
Reviewer 2 Report
Detailed comments:
Line 12: Please change "increasing scholars" as it sounds rather strange.
Line 17: Do you mean certain scenarios?
Line 17: Please describe what low-channel lidar refers to?
Line 18: Why you include (slice-based) here? There are other novel non-slice-based methods as well?
Line 19: "It" refers to your novel method? If yes please rework this part of the abstract to highlight your method consisted of two parts.
Line 19: What about the semantic segmentation or the other part?
Line 32: What do you mean by traffic perception here? This needs to be defined as well.
Line 35: Please remove the word mortal danger and replace it with a more appropriate alternative.
Line 39: Please define what is microscale traffic data?
Line 40: ... by fade areas "and" sensor faults ...
Line 44: Please clarify what do you mean by brightness.
Line 47: The "L" in Light detection and ... should be lower case.
Line 49: a lidar sensor can have 360 coverage, but not all of them scan the 360 degrees.
Line 50: Please name the traffic flow parameters.
Line 52: please update the unsupervised scenarios to unseen scenarios as "unsupervised" can confuse the readers.
Line 53: Please change the "increasing scholars".
Lines 55 and 56: This sentence is very similar to the sentence used in the abstract. Please make sure to paraphrase one of them.
Line 57 to 61: The given definitions for semantic segmentation and instant segmentation are not accurate. Please update these. In semantic segmentation, the label for each pixel or point is determined while similar objects are group together. In instance segmentation, similar objects are detected and distinguished.
Line 77: Please update the sentence for grammar and readability.
Line 78: Please provide a more valid reason why the edge-based methods are not suitable for your data.
Line 85: The studies within 16 to 18 needs to be discussed. Please briefly describe them.
Line 87: Please update the sentence for grammar.
Line 93: The generalization that states graph-based methods can ... is not true. Please remove this sentence.
Line 96: Please clarify what do you mean by handcrafted methods and learning-based methods? Classifications are either supervised or unsupervised methods (which include classification base on defining criteria).
Line 98 to 105: These descriptions are not accurate at all. PointNet, 3DCNNs, and other variations do not extract features. They learn to classify data without any feature extraction process. Please update these definitions.
Line 102: What is an irregular point cloud? Do you unstructured point cloud?
Line 111: Please define infrastructure-based lidar? what are the differences between terrestrial, mobile, and airborne lidar sensors and roadside lidar and infrastructure-based lidar?
Line 117: Where are Figures?
Line 122: why the color here is red?
As figures are not provided with the text, it is rather impossible to review this paper further. I will stop and look for a more complete version of the paper to review.
Author Response
Dear Reviewer:
We are sorry for missing figures in the original manuscript. We think that the figure format we used in the original manuscript caused the figures cannot display correctly. We have revised and converted the figure format in the revision version.
We wish to express our very deep appreciation, and the appreciation of all of us, to your patience, great efforts and suggestions for our manuscript under missing figures. The comments and suggestions are valuable and very helpful for revising and improving our paper, as well as the important guiding to our researches.
The attachment is a point-to-point response to your comments and the responses are in red. The modification marked in red in revised version.
We tried our best to improve the manuscript and made some changes in the manuscript. These changes will not influence the content and framework of the paper. And here we did not list the changes but marked in red in revised paper. We appreciate for your warm work earnestly, and hope that the correction will meet with approval. Thank you for your time and patience. I look forward to receiving your letter.
Once again, we would like to thank you for the constructive comments and suggestions. Please feel free to contact us with any questions. We are looking forward to your reply.
Yours sincerely,
Authors
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The manuscript can be accepted in present form
