Optical Flow-Based Algorithms for Real-Time Awareness of Hazardous Events

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper proposed optical flow-based automated approaches for a multitude of situation awareness and event alerting challenges. However, the paper contains some theoretical errors. The specific comments are as follows. 
1.    There are contradictions among the figures, formulas, and corresponding descriptions in this paper, e.g., the green arrow in Figure 1, the relationship between GLORIA and SOFIA in Figure 1, etc.
2.    Add a section specifically dedicated to introducing the relevant research work. Moreover， the differences from related methods should be highlighted.
3.    In the paper, there are formulas derived from other research works. It is better to focus only on the contribution without repeating obvious things.
4.    To validate the effectiveness of the method, a comparison with baseline methods is essential. Additionally, ablation experiments should be carried out.
5.    To demonstrate the effectiveness, this method should be compared with other state-of-the-art methods.
6.    The references have some errors, e.g., [3], [87], [93], etc.

Author Response

Please see the file attached

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a comprehensive methodological framework describing two proprietary optical-flow algorithms,SOFIA and GLORIA, and their application to the detection of hazardous events such as convulsive seizures, apnea, falls, and explosions, as well as for object tracking and video stabilization. The topic is highly relevant and the technical foundation is solid; however, the paper would benefit from several improvements to enhance clarity, focus, and scientific depth. Please see below my comments

1. Figure 1 and the accompanying description could be expanded to clearly illustrate the full data-processing pipeline — from video input to the detection module output — and to explain how SOFIA and GLORIA interact within the system architecture.
2. Sections 2.1–2.8 are currently very lengthy and may be condensed, allowing the authors to place stronger emphasis on experimental validation and performance assessment rather than extended mathematical exposition.
3. Sections on apnea, falls, and explosions rely heavily on earlier publications. Clarify what is new in this work compared to previous studies.
4. Add quantitative results such as accuracy, sensivity, specifity , latency and others. to support claims of robustness and real-time performance.
5. Include at least one benchmark against established optical flow or deep-learning methods (e.g., Horn–Schunck, Lucas–Kanade, FlowNet2, RAFT).
6. Reduce self-citations and add recent references  (from the past five years) to reflect current developments in optical-flow and event-detection research.

Author Response

Please see the file attached

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have replied to my comments.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed all of my comments and incorporated the revisions into the original manuscript. From my perspective, the manuscript is now ready for publication.