Classification of Typical Static Objects in Road Scenes Based on LO-Net

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper presents a novel deep learning approach, LO-Net, for enhancing point cloud classification accuracy in mobile LiDAR data, demonstrating significant improvements in object identification in road scenes, particularly with the custom Road9 dataset. While the amount of work is commendable, there are some issues in the article that need to be addressed:

1. The core problem definition in the introduction is not concise enough. A concise summary of the core problem and the article's motivation is needed.

 

2. Related work should not merely list relevant algorithms. Please condense this section and discuss the methods of this paper.

 

3. The experimental discussion is confusing and repetitive. It is recommended to distinguish the results corresponding to the datasets.

4. In Table.6, I suggest changing the table to a visual presentation style. It’s so hard to read

 

5. There are some minor errors in the paper, like the wrong figure label (see Fig. 10)

Comments on the Quality of English Language

no more comments

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Appreciate your efforts in introducing this study related to LiDAR object detection and classification. Though work presented is very interesting, manuscript lack in various key aspects to convey the actual scope and contribution. Research objective and contribution must be discussed clearly also methodology and experimental section needs to modified. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Minor English correction needed at some places. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Frist of all, I would like to thank the authors for their efforts to produce a well-written manuscript that provides very informative detials about the work presented to the reader. Please, consider my minor comments below:

line 16 and 19 : write the full form of the abbreviations SA and J-PSPP

line 25: what is ROAD9 dataset. The reader is not yet famililar with the naming conventions you define in the paper for the dataset. Please clarify.

Please, use past tense in the conclusion Section. For example, line 604: This paper introduces. Changes it to introduced. 

Line 131: you define the SA abbreviaiton. It should appear earlier in line 109. Please modify.

Best of luck.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Appreciate your efforts in modifying the manuscript accordingly. Though current version of manuscript appears to improved in terms of the scope of proposed research work, its still far from conveying actual research contribution, new findings and enhancements. 

As stated in title and abstract, proposed approach does not discuss any challenges of existing approach and nor any significant enhancement. Moreover, it is not clear that why title of manuscript specific focus on only road scene ? While in the experimental analysis and discussion emphasis given more on normal object detection without specific to road scene. Title needs to clearly mention to convey actual scope. Object classification specific to road scene should include classification of objects such as car, truck, motor vehicle, bicycle, pedestrian, traffic light,  bus shelter, and so on.

Moreover, existing approach limitations on partial objects may not be appropriate. Clear illustration with sample images should be included. Objects shown in the Table 6, reality doesnt looks like partial data. Partial data should be clearly defined according to natural scenarios. 

In the Table 6,  its surprising to see none of method able to detect the traffic light though point cloud is almost complete.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx