A Real-Time Shipping Container Accident Inference System Monitoring the Alignment State of Shipping Containers in Edge Environments

Round 1

Reviewer 1 Report

The authors report a shipping container alignment detection method using yolov4 and edge processor. The paper has some interesting aspects, but the comparison experiments are weak. Solid experimental comparison with other methods are suggested to be added.

- Yolov4 is out of date since yolov8 has been proposed.

- There are no comparisons with other STOA methods.

 

Author Response

Title: Real-time Shipping Container Accident Inference System through Monitoring Alignment State of Shipping Container in Edge Environments

Authors: Se-Yeong Oh, Jun-Ho Jeong, Sang-Woo Kim, Young-Uk Seo, Joosang Youn* *Correspondence

We really appreciate your time and valuable comments. We have carefully revised our paper taking into consideration your comments and suggestions. The following attached file are detailed responses to your comments. (C: Comment, A: Answer). Please note that modifications are marked as bold font in the revised manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript developed a Real-time Shipping Container Accident Inference System by Monitoring the Alignment State of Shipping Container in Edge Environments. The proposed method focuses on exploring the solution for intelligent port safety services among various intelligent port services. Looking at the recent solution for intelligent port safety services in the port yard, the solutions applying object recognition technology are being actively developed to prevent safety accidents.

 

Overall, the topic of this manuscript is interesting, and the proposed algorithm is evaluated by real-world camera datasets containing shipping containers, which can show the strong generalization ability of the system for the target task.

 

My comments are major include:

1. Considering the possible camera distortion, do the definitions of the “safe zone”, “caution zone” and “dangerous zone” from ISO1611 still work to estimate the container alignment?

2. Although the used metrics (Precision, Recall and F1-score) are fine to evaluate the performance of the proposed method, it is more suggested to implement the common-used average precision (along iou from 0.5 to 0.9) to test the algorithm.

3. Besides YOLO, there are a series of modern object detection methods which can also achieve promising performance in different object detection tasks, such as [1-5]. The related work section will be better if the authors provide a more comprehensive review of the related research.

Ref.

[1] Jiwoong Choi, Dayoung Chun, Hyun Kim, and Hyuk-Jae Lee. Gaussian yolov3: An accurate and fast object detector using localization uncertainty for autonomous driving. In Proceedings of the IEEE International Conference on Computer Vision, 2019.

[2] Yihui He, Chenchen Zhu, Jianren Wang, Marios Savvides, and Xiangyu Zhang. Bounding box regression with uncertainty for accurate object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019.

[3] Xiyang Dai, Yinpeng Chen, Jianwei Yang, Pengchuan Zhang, Lu Yuan, and Lei Zhang. Dynamic detr: End-to-end object detection with dynamic attention. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021.

[4] Le Yang, Ziwei Zheng, Jian Wang, Shiji Song, Gao Huang, and Fan Li. An Adaptive Object Detection System based on Early-exit Neural Networks. In IEEE Transactions on Cognitive and Developmental Systems, 2023.

[5] Xiyang Dai, Yinpeng Chen, Bin Xiao, Dongdong Chen, Mengchen Liu, Lu Yuan, and Lei Zhang. Dynamic head: Unifying object detection heads with attentions. In Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition, 2021.

Author Response

Response to Reviewer 2’s Comments
Title: Real-time Shipping Container Accident Inference System through Monitoring

Alignment State of Shipping Container in Edge Environments

Authors: Se-Yeong Oh, Jun-Ho Jeong, Sang-Woo Kim, Young-Uk Seo, Joosang Youn* *Correspondence

We really appreciate your time and valuable comments. We have carefully revised our paper taking into consideration your comments and suggestions. The following attached file are detailed responses to your comments. (C: Comment, A: Answer). Please note that modifications are marked as bold font in the revised manuscript.

Author Response File: Author Response.pdf

Reviewer 3 Report

Real-time Shipping Container Accident Inference System through Monitoring Alignment State of Shipping Container in Edge Environments

artificial intelligence(AI) can be defined to use it frequently throughout.

Avoid such lengthy statements, “In this paper,

we propose a real-time shipping container accident inference edge system that can monitor the ship

container loading status in real time and analyze the loading status of ship containers from a safety

point of view in order to prevent safety accidents related to misaligned loading of ship containers in

advance.”

Can be divided further.

Last line of abstract: “In this paper, we consider the port safety scenario, since

inference accuracy is more important than inference speed. Therefore, we apply the YOLOv4 network

to the algorithm of the inference model.”

It can be removed, as not in connection or flow:

Use some of recent works alos towards image segmentation using ML, suggestions could be, EnsembleNet: A hybrid approach for vehicle detection and estimation of traffic density based on faster R-CNN and YOLO models, Convolution network model based leaf disease detection using augmentation techniques

Table 1: and other algorithm[13], what is this other algorithm?

Figure 2, needs to be improved, as no need to show back network, hub switches, firewall,l2 etc.

Readers would be more interested to know, which edge server is used and was that physical or some service?

Do some labelling to figure 4,5, and 6 for readability.

Figure 8, meaasge str of which protocol or it is designed by authors?

Table 4, how is data balance is taken care in images?

Add improvments in conclusion as well.

 

 

 

 

 

 

 

Moderate improvements

Author Response

Response to Reviewer 3’s Comments
Title: Real-time Shipping Container Accident Inference System through Monitoring Alignment State of Shipping Container in Edge Environments

Authors: Se-Yeong Oh, Jun-Ho Jeong, Sang-Woo Kim, Young-Uk Seo, Joosang Youn* *Correspondence

We really appreciate your time and valuable comments. We have carefully revised our paper taking into consideration your comments and suggestions. The following attached file are detailed responses to your comments. (C: Comment, A: Answer). Please note that modifications are marked as bold font in the revised manuscript.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed my concerns, therefore it can be accepted as it is.