Next Article in Journal
Long-Term Investigation of Aerosols in the Urmia Lake Region in the Middle East by Ground-Based and Satellite Data in 2000–2021
Next Article in Special Issue
Deep-Learning-Based Three-Dimensional Detection of Individual Wave-Dissipating Blocks from As-Built Point Clouds Measured by UAV Photogrammetry and Multibeam Echo-Sounder
Previous Article in Journal
Image-Based Obstacle Detection Methods for the Safe Navigation of Unmanned Vehicles: A Review
Previous Article in Special Issue
2D&3DHNet for 3D Object Classification in LiDAR Point Cloud
 
 
Article
Peer-Review Record

DGPolarNet: Dynamic Graph Convolution Network for LiDAR Point Cloud Semantic Segmentation on Polar BEV

Remote Sens. 2022, 14(15), 3825; https://doi.org/10.3390/rs14153825
by Wei Song 1, Zhen Liu 1, Ying Guo 1,*, Su Sun 2, Guidong Zu 3 and Maozhen Li 4
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3:
Remote Sens. 2022, 14(15), 3825; https://doi.org/10.3390/rs14153825
Submission received: 20 June 2022 / Revised: 29 July 2022 / Accepted: 4 August 2022 / Published: 8 August 2022

Round 1

Reviewer 1 Report

The paper proposes a method for point cloud segmentation with the contribution of extracting the local features mainly edges in the enchantment of previously developed models. The work is well explained already and includes a comprehensive discussion. 

I have some minor comments as follows;

*Figure 1 .. Dynamin -->Dynamic

*Figure 4 needs zoom to the objects, difficult to interpret

*better to include the results in figures in the supplementary dataset

*better to discuss the reasons for FP results

*The network is better to be open to the public when the paper is published.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper proposes Dynamic Graph Convolution Neural Network for LiDAR Point Cloud Semantic Segmentation.  The system was tested on the Semantic KITTI dataset, and the segmentation accuracy reached 56.5%. The paper has significant scientific contributions. However, the paper should be aupdated with 2021 and 2022 papers                         

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

1. In the introduction, the problems that need to be solved by analyzing the existing related conditions are not directly supported in the subsequent contributions and experimental analysis.

2. In Section 2, the logic of the upper and lower paragraphs is not clear and direct;

3. Figure2 is the core module of the paper. After listing, please explain the diagram in detail.

4. Is it possible to add repetitive experiments and ablation experiments?

5. Remote Sensing is an excellent journal. When submitting, pay attention to learning and citing related papers.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Figure3 is the core module of the paper. thought authors added content in recently verision manuscirpt, it is still difficult to understand to ordinary scientific researchers. Please give an example and introduce it module by module so that more readers can understand it easily. for example, every input and output meaning, 1*32*32, 1*32*32*19,each number meaning.

A good manuscirpt is to write a profound theory so that others can easily understand it.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Back to TopTop