Multi-Object Detection for Inland Ship Situation Awareness Based on Few-Shot Learning
Round 1
Reviewer 1 Report
This paper proposed a novel method for USVs multi-object detection and tracking. Overall, this topic has some potential and could be applicable. Some minor modifications should be made before publication.
1) Currently, there are some integrated frameworks for detection and tracking: integrated inertial-lidar-based map matching localization for varying environments, hydro-3d: hybrid object detection and tracking for cooperative perception using 3d lidar, an automated driving systems data acquisition and analytics platform. Please discuss this pipeline and compare them with your proposed framework to highlight your innovation.
2) For the module of tracking, you adopted the Kalman filter, actually, equation (7) to (9) is a traditional KF. In: automated vehicle sideslip angle estimation considering signal measurement characteristic. Based on the above work, the advantages of KF should be discussed in detail.
3) When you compare with other SOTA methods, the parameter size and inference time should be listed.
4) Figure 2 can be deleted as it is only the original framework not your proposed methods.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
In this study, a new method for ship detection is proposed based on the YOLOv5 algorithm. The subject of study is popular and interesting. However, there is no tracking part in the study. The study should be developed in accordance with the following recommendations.
1) Why didn't you use more recent versions of YOLO?
2) “4.3. Transfer learning training strategy” is not referenced in this section. Authors are expected to cite studies in the literature for information they did not produce.
3) Comparing your results with more recent YOLO versions will be beneficial for the effectiveness of your study. (YOLOv7, YOLOV8 etc.). Compared YOLO versions are out of date.
4) There is no evaluation of object tracking in the results. In the title, it is stated that it includes tracking along with detection. The tracking application should be added in the study or the tracking statement should be removed from the title.
5) Training and testing speeds of the algorithms should be evaluated. The claim that the proposed method is fast must be scientifically proven.
6) The proposed method should be tested on public datasets.
Minor editing of English language required
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
In my opinion, the manuscript is acceptable as is.
