Pre-Inpainting Convolutional Skip Triple Attention Segmentation Network for AGV Lane Detection in Overexposure Environment
Round 1
Reviewer 1 Report
The authors presented in the article the problem of lane detection, which is important for the visual navigation of AGV vehicles. Currently, deep learning based image segmentation is very popular in lane detection. However, the lane detection accuracy is severely degraded due to overexposed images of the lane by flash or sunlight. Therefore, the authors proposed a lane detection method combining image painting and image segmentation for incomplete image segmentation of the lane extraction under intense lighting.
The article is well written. The authors presented a concise review of the literature showing the achievements of other authors in this field. They described the proposed method and presented the results of the conducted experimental studies. The obtained results were compared with the results obtained by other scientists. The article ends with conclusions summarizing the work. I have no comments regarding the article.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
This study examines Pre-inpainting Convolutional Skip Triple Attention Segmentation Network for AGV Lane Detection in Overexposure Environment. The subject matter of this manuscript fits the journal's scope, and the information included in the manuscript seems not to have been published in any other publication so far. However, it seems difficult to adequately evaluate the value of this study because the explanation of the significance of the study, the description of the interpretation and usefulness of the results obtained by the analysis, and the explanation of the model are insufficient. I would like to ask the authors to consider responding to the following comments:
(1) Abstract didn’t summarise all the key findings of the manuscript
(2) Would you explicitly specify the novelty of your work? What progress against the most recent state-of-the-art similar studies was made?
(3) The Introduction section should be improved. It should be dedicated to presenting a critical analysis of state-of-the-art related work to justify the study's objective. In addition, critical comments should be made on the results of the cited works.
(4) The main objective of the work must be written in a more precise and concise way at the end of the introduction section.
(5) Please carefully check recent literature and discuss/cite as you see fit, and update your reference list such as; Public Preferences Towards Car Sharing Service: The Case of Djibouti. Prediction of electric vehicle charging duration time using ensemble machine learning algorithm and Shapley additive explanations. A comparative performance of machine learning algorithm to predict electric vehicles energy consumption: A path towards sustainability. Grey wolf optimizer-based machine learning algorithm to predict electric vehicle charging duration time
(6) The provided experiment setting was well-conducted in this manuscript, but the comparative methods (ERFNet, CCNet, PP-LiteSeg, Proposed) were still not convincing. Further comparison experiments with state-of-the-art methods are required to provide the outperformance of the proposed method.
(7) The conclusion section provides a lack of contributions to this manuscript. Provide the key features, merits, and limitations of the proposed approach to clarify the precise boundary of the algorithms. The implication of the proposed method is also required.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
1. "The complexity of scenes of lane detection is low. But the system has high requirements for real-time detection. Therefore, the structure of MAE can be simplified appropriately, so as to ensure the quality of image restorement and accelerate the speed of image restorement. In the proposed method, the number of layers of Transformer encoder blocks is set to 6. The number of layers of Transformer decoder blocks is set to 4." Why 4?
2. Fig 7 and fig 8 can improve comments
3. Table 1 and 2 can improve comments
4. Comment on other options outside detection by lines for AVGs
5. Conclusions- are poor. Improve
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The author has addressed all my previous comments adequately
