Concealed Object Detection and Recognition System Based on Millimeter Wave FMCW Radar
Round 1
Reviewer 1 Report
Overall, it is a good paper, but needs improvement. The idea of mm-wave MIMO radar is a relatively new concept.
Several grammatical errors and incomplete sentences are noted and these must be corrected. For example, this is not a complete sentence: “Due to the frequent occurrence of terrorist activities in recent years, and mostly for airports, railway stations, subway and other crowded public places [1].”
Authors state that: “It can penetrate the surface to image the hidden target.” Actually, this is not strictly true as it depends upon the material. Also, it has limitations in that it can only penetrate to certain extent. Depth of penetration must be clearly discussed.
Delete names of researchers. Instead of saying, for example, “In 2020, Josiah Wayland Smith used MIMO-ISAR technology to reduce scanning time in a near-field millimeter-wave imaging system [11]. In 2021, Mailad Rezaei used dual-polarization antennas to improve millimeter-wave imaging system [12].”, it is better to say “In 2020, MIMO-ISAR technology was used to reduce scanning time in a near-field millimeter-wave imaging system [11]. In 2021, dual-polarization antennas were employed to improve millimeter-wave imaging system [12].”
All the images in Figure 11 pertain to each item inside an empty box. This is an easy problem, almost like free space. Have you tried to detect objects concealed inside thick coat pockets, as this is a more plausible and realistic scenario? I suggest you also take additional data and show similar images when the items are hidden in a thick coat pocket. Otherwise, this paper is not useful.
Some of the figures are too small and not readable. They must be increased in size so that the letters are readable.
The probing wavelength used is very small, ~3.9 mm. In a realistic scenario, target containers and humans carrying targets can sway and move by more than the wavelength. Discuss the effects and how the algorithm corrects for such errors.
Author Response
Dear Reviewer 1,
Thank you for your report.
We would also like to thank you, for taking the time to read our paper and for your important comments and remarks.
A detailed reply to your comments is given below.
Yours sincerely
Kai Zhang
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The manuscript has some potential for journal publication, however at present it looks like not fully completed technical report. The structure and writing should be completely reworked, but as interesting fusion of Deep learning for MIMO radar signal processing is suggested, my recommendation is not to reject the paper but to allow major revision.
The approach is very simple from FMCW radar viewpoint, and more interesting from DL, but the later part is not carefully explained. Thus, the second part should be emphasized. Overall, the novelty should be discussed – what is new or novel regarding designed FMCW sensor, imaging approach, and DL approach for object recognition with detailed performance analysis and recognition accuracy statistics, with reference to state-of-the-art.
Such issues should be reworked:
1. FMCW overview is very limited and should be strengthened. It is widely used nowadays in health monitoring, using not only amplitude, but also phase (difference between concequitive/neighbouing sweeps). See for example these several pieces:
- Rong Y, Dutta A, Chiriyath A, Bliss DW. Motion-Tolerant Non-Contact Heart-Rate Measurements from Radar Sensor Fusion. Sensors. 2021; 21(5):1774. https://doi.org/10.3390/s21051774
- Turppa E, Kortelainen JM, Antropov O, Kiuru T. Vital Sign Monitoring Using FMCW Radar in Various Sleeping Scenarios. Sensors. 2020; 20(22):6505. https://doi.org/10.3390/s20226505
- Baek S, Jung Y, Lee S. Signal Expansion Method in Indoor FMCW Radar Systems for Improving Range Resolution. Sensors. 2021; 21(12):4226. https://doi.org/10.3390/s21124226
- and many others.
2. There are terminology issues. SAR appears in the abstract without disambiguation, but also X- and Y- directions are mentioned. In SAR imaging, “azimuth” and “range” are used. Morover, as Doppler effects are not used, like in regular SAR image formation (from airlplanes, drones or satellites), the terminology should be revised. To my understanding, the authors are dealing with MIMO radar, which is not classical SAR (single-channel SAR) principle, but just MIMO radar as such, like in paper by Turppa et al above. Please note that there is also “MIMO SAR” radars, which uses both sensor motion in the azimuth direction, and multichannel MIMO, and that is much more advanced construct https://www.researchgate.net/publication/257008458_MIMO_SAR_techniques_and_trades
The authors are not fully wrong per se, but it is important to be very clear here in terminology, as using SAR, MIMO, and MIMO SAR interchangeably can cause confusion.
3. The parts between FMCW and DL should be separated in Introduction as subsections, to introduce and explain the need for DL in this context. Overall, existing gaps in the literature and need/necessity of this research should be clearly explained.
4. In Materials and Methods, the authors start from explaining basic FMCW principle. This is unclear and misplaced. If they want, they can introduce a separate “theory” subsection in Introduction, however, the authors do not tell anything new regarding FMCW principles, so it can’t belong to “Methods”. Methods should concentrate on DL, while explaining peculiarities of input signals. This is partly done in section 2.3, however, this can also belong to Introduction, or to separate “Image formation” section.
5. Discussion has items that can/should belong to Conclusions, but discussion is missing as such. As a workaround, authors could have “Results and Discussion” and “Conclusions” sections. Also, Conclusion section is missing completely.
6. Demonstration is very limited. What is the need for “object recognition” then? What happens to “tilted” objects? What are performance limits? Generally, the analysis can be improved and more experiments done, improving the Discussion in the paper.
7. Big issues are in the language. Very long sentences with several ideas should be split, e.g. 461-463. Also, English should be revised, like “image are put into the…” ?, “increase the recognition of the image to human eye” - > perhaps “improve visual recognition” , “two round measurements were measured”?, “correspondently large” ?
Author Response
Dear Reviewer 2,
Thank you for your report.
We would also like to thank you, for taking the time to read our paper and for your important comments and remarks.
A detailed reply to your comments is given below.
Yours sincerely
Kai Zhang
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
My concerns have been adequately addressed by the authors.
