Smartphone User Identification/Authentication Using Accelerometer and Gyroscope Data
Round 1
Reviewer 1 Report
The authors present a study on Smartphone User Identification Using Accelerometer Data. The article uses well-known techniques, and there is no significant contribution to the existing scientific knowledge. However, the review has the following critical concerns regarding the current version of the article.
Abstract
The article’s abstract must present a clear view of the problem statement and purpose of the study. However, this article lacks a problem statement and objective in the abstract. The authors must rewrite the abstract by considering the following keypoints in sequence.
i. Background statement.
ii. Research gap
iii. Proposed solution
iv. Evaluation matrices
v. Significance of the proposed solution
Introduction
The article's introduction is normally an extensive version of the abstract and presents a list of contributions followed by the article's organization. However, there is no sequence in the current version, and the introduction does not correspond to the abstract. There are several acronyms used but not defined, such as HMOG, LSTM, GPS, etc. There is no list of contributions. Why the authors presented information in bullet form? It is better to write logical paragraphs with proper citations rather than adding bullet points in the introduction.
Related Work
Related work is fine. However, a table summarizing related work would be better from the reader's point of view.
Materials and Methods
In materials and methods, the font size of the tables must be increased. There are several missing references to the tables, such as captions of Figures 4, 5, 6, and 7 have "??". It is recommended to draw these figures with the help of Python or MATLAB to improve the quality of the figures.
One critical point is the lack of description of the methods used in the study. And what is the author's significant contribution rather than using existing techniques?
Results
Comparison with state-of-the-art models is missing current table 7 for comparison is insufficient. Moreover, the results presented show the performance of existing models. What about the author's proposed method? After studying sections 3 and 4, it looks like the authors have not proposed anything. They just evaluated existing models.
Conclusion
The conclusion must be coherent with the study presented, and authors must avoid bullet points in the conclusion also.
In summary, the article just presents the evaluation of existing ML classifiers on the publically available dataset without contributing to scientific knowledge.
Some grammatical errors must be improved, and the article lacks coherence and cohesion.
Author Response
The authors present a study on Smartphone User Identification Using Accelerometer Data. The article uses well-known techniques, and there is no significant contribution to the existing scientific knowledge. However, the review has the following critical concerns regarding the current version of the article.
Authors' Response: Indeed, we acknowledge the reviewer's observation that the initial contribution of our study was limited to user identification using accelerometer data alone. We appreciate the reviewer's feedback, which prompted us to enhance our research by addressing the identified gap. Consequently, we conducted additional experiments focusing on user authentication and incorporating gyroscope data alongside accelerometer data. By combining the data from both sensors and utilizing appropriate methodologies, we achieved nearly perfect authentication results. We have dedicated a specific subsection in the results section to present and discuss these new contributions. Furthermore, we have added another subsection that summarizes both the old and new contributions, clearly highlighting these enhancements in red. Therefore, we had to change the title of the paper to be “Smartphone User Identification/Authentication Using Accelerometer and Gyroscope Data”. We believe that these additions significantly strengthen the value and impact of our study. We appreciate the reviewer's insights, which have helped us improve the significance of our research.
Abstract
The article’s abstract must present a clear view of the problem statement and purpose of the study. However, this article lacks a problem statement and objective in the abstract. The authors must rewrite the abstract by considering the following keypoints in sequence.
- Background statement.
- Research gap
iii. Proposed solution
- Evaluation matrices
- Significance of the proposed solution
Authors' Response: We appreciate the reviewer's comment, and we have revised the abstract to address their key points.
Introduction
The article's introduction is normally an extensive version of the abstract and presents a list of contributions followed by the article's organization. However, there is no sequence in the current version, and the introduction does not correspond to the abstract. There are several acronyms used but not defined, such as HMOG, LSTM, GPS, etc. There is no list of contributions. Why the authors presented information in bullet form? It is better to write logical paragraphs with proper citations rather than adding bullet points in the introduction.
Authors' Response: We appreciate the reviewer's feedback and have made significant revisions to the introduction based on their comments. The introduction now provides a comprehensive overview of the study's contributions and incorporates the organization of the article. Furthermore, all acronyms used, including HMOG, LSTM, and GPS, have been properly defined to ensure clarity for readers.
To address the reviewer's concern about the bullet-point format, we have transformed the paragraph into logical paragraphs with proper citations. This change enhances the readability and flow of the introduction, making it more coherent and aligned with standard academic writing practices.
Additionally, we have taken note of the reviewer's suggestion to provide a specific list of contributions. Consequently, we have included a dedicated section at the end of the Results section to succinctly outline the key contributions made in this study.
These revisions significantly improve the structure and clarity of the introduction, ensuring that readers can better understand the research objectives, contributions, and overall organization of the article. We appreciate the reviewer's valuable insights, which have enabled us to enhance the quality of our work.
Related Work
Related work is fine. However, a table summarizing related work would be better from the reader's point of view.
Authors Response: We appreciate the reviewer's comment and acted based on it by creating a new table (1), showing summary of some literature: number of users, feature selection methods, classifiers, and the performance metrics used.
Materials and Methods
In materials and methods, the font size of the tables must be increased. There are several missing references to the tables, such as captions of Figures 4, 5, 6, and 7 have "??". It is recommended to draw these figures with the help of Python or MATLAB to improve the quality of the figures.
Authors' Response: We appreciate the reviewer's feedback and have made the necessary adjustments based on their comments. The font size of the tables has been increased to improve readability. Additionally, the missing references and captions for Figures 4, 5, 6, and 7 have been rectified to provide proper labeling and context. To enhance the quality of the figures, we have redrawn them using Python and exported them as PDF files.
One critical point is the lack of description of the methods used in the study. And what is the author's significant contribution rather than using existing techniques?
Authors' Response: Regarding the lack of description of the methods used in the study, we have addressed this concern by expanding the Materials and Methods section to provide a comprehensive explanation of the techniques employed. This includes detailed descriptions of the data preprocessing, feature extraction, feature selection, handling class imbalance, and the various machine learning classifiers utilized.
As for the significant contribution of the study, we have taken the reviewer's suggestion into account and now provide a specific list of contributions at the end of the Results section. This ensures clarity and highlights the unique contributions made beyond the utilization of existing techniques. We appreciate the reviewer's valuable input, which has helped us improve the presentation, quality, and clarity of our work.
Results
Comparison with state-of-the-art models is missing current table 7 for comparison is insufficient. Moreover, the results presented show the performance of existing models. What about the author's proposed method? After studying sections 3 and 4, it looks like the authors have not proposed anything. They just evaluated existing models.
Authors' Response: We appreciate the reviewer's feedback and have made significant improvements based on their comments. Firstly, we have enhanced the comparison by adding two more state-of-the-art works to Table 7, providing a more comprehensive analysis.
Regarding the perceived lack of a proposed method, we acknowledge that the initial contribution was weak, primarily focused on evaluating existing models. However, we have taken the reviewer's comments seriously and conducted additional experiments to address this limitation. Specifically, we have expanded our research to include user authentication and incorporated the gyroscope signal alongside accelerometer data. The combination of these sensors has resulted in almost perfect user authentication, which significantly enhances our contribution.
To clarify the evolution of our contribution, we have added a dedicated section in the paper that details both the old and new contributions. This section explicitly outlines how our study has progressed from evaluating existing models to proposing and demonstrating the effectiveness of the combined accelerometer and gyroscope approach for user authentication.
We sincerely appreciate the reviewer's valuable insights, which have motivated us to enhance the significance and impact of our research.
Conclusion
The conclusion must be coherent with the study presented, and authors must avoid bullet points in the conclusion also.
Authors' Response: We acknowledge the reviewer's comment regarding the coherence of the conclusion section and the use of bullet points. In response to this feedback, we have thoroughly revised the conclusion section, ensuring that it aligns coherently with the study presented. We have refrained from using bullet points and instead have crafted well-structured paragraphs that summarize the key findings and implications of our research. By addressing the reviewer's concerns, we have improved the overall flow and readability of the conclusion section, allowing for a more comprehensive and effective closure to our study. We appreciate the reviewer's input, which has helped us enhance the quality and coherence of our paper.
In summary, the article just presents the evaluation of existing ML classifiers on the publically available dataset without contributing to scientific knowledge.
Authors' Response: We appreciate the reviewer's feedback and have taken steps to address their concerns. Initially, our study focused on the evaluation of existing machine learning classifiers on a publicly available dataset, which may have appeared to lack a significant contribution to scientific knowledge. However, we have since conducted additional experiments that have substantially enhanced the value of our research.
By incorporating more experiments on user authentication and leveraging additional data, specifically by utilizing the gyroscope signal in conjunction with accelerometer data, we have achieved nearly perfect user authentication results. This significant improvement in our findings establishes a more substantial contribution to the scientific knowledge in this field. To provide clarity on the evolution of our contribution, we have included a dedicated section in the paper that outlines both the old and new contributions of our study. This section explicitly highlights the enhancements made, particularly in the realm of user authentication.
Comments on the Quality of English Language
Some grammatical errors must be improved, and the article lacks coherence and cohesion.
Authors' Response: We appreciate the reviewer's feedback regarding grammatical errors, coherence, and cohesion. To address these concerns, we have taken the necessary steps to improve the language quality of the manuscript. A native English speaker has reviewed the article and corrected any identified grammatical errors.
Regarding coherence and cohesion, we acknowledge that the earlier version of the manuscript may have lacked consistency due to discussions encompassing all smartphone sensors while ultimately focusing on only one sensor. Additionally, the related work section primarily discussed authentication, while our study primarily focused on user identification. However, we have made significant improvements to address these issues. Specifically, we have added the authentication component to our study, which has contributed to a more coherent narrative. Additionally, we have incorporated the gyroscope data as an additional signal, further enhancing the overall cohesion of our manuscript.
By making these revisions, we believe that the manuscript has become more coherent and cohesive, providing a clearer and more focused presentation of our research. We appreciate the reviewer's feedback, which has helped us enhance the quality and consistency of our work.
Reviewer 2 Report
A) General remarks
The research presents in this paper a very interesting topic, as well as results that are of wider significance when it comes to the application of vibration measurements in person identification. The paper is, however, not always concise and clear. The literature in the paper is adequately cited.
1. In the case of literature, all the cited references are relevant to the research. No unnecessary self-citations were detected.
2. The English language used is without major mistakes, however, some sentences need to be checked by a native speaker.
3. The abstract is well written. The role of the abstract is to give a basic overview of the paper. In this case, the abstract gives a good introduction to the paper without specific data and is very informative even for those none-familiar with the topic reader. However, the novelty aspects of the paper are not presented.
4. The reviewer would also suggest not using the abbreviation in the abstract but the full names and introducing the abbreviation the first time it appears in general text.
5. It seems that the authors use sometimes impersonal and sometimes personal forms. It is more common to use impersonal forms in scientific papers. Use only one convention in the whole article.
6. The introduction is mostly well-written and follows all the rules of the proper instruction on the topic. However,
a. In general, the introduction is quite basic and is more typical for a magazine publication than a scientific paper.
b. The last paragraph where the aim, scope and novelty are to be presented is quite short and the novelty aspect is weak or not existing.
7. Chapter 2 is well written
8. Material and methods:
a. no information on the sensor's parameters.
b. No photos of the test conditions would clarify how the test was conducted. No connection between the human orientation with the orientation of the phone. No explanation of how X,Y,Z axis on Fig. 2 are situated during the test. Thus this figure is meaningless.
c. Fig 2 and 3 are basic, well know information that brings nothing to the reader except the one unfamiliar with statistical or signal analysis.
9. The paper does provide proper conclusions. However, again the novelty aspect must be pointed out strongly.
10. The biggest problem of the paper is that the authors are writing about some solutions like using an accelerometer, gyroscope and magnetometer but after only using an accelerometer. Similarly with the algorithms. This makes the paper not truly thought through and some things difficult to follow.
B) Item remarks
Some figures are out of focus probably due to pdf compression. Make sure that this is not the case in the final version
Fig .9,10, are hardly readable- please enlarge them and make sure that axis captions and legends are visible.
Tab. 7 – comment again on the personal/impersonal form. It looks better to write in impersonal form like in tab 7 “ours” looks strange but if changed to “Authors” or “authors solution” looks much more professional.
C) Conclusions:
The biggest problem of the article is also the clear presentation of the novelty of the research topic and in many places confusing structure or claims that are not followed later in the test or results presentation.
Thus, the reviewer suggests the major corrections mentioned previously and asks the authors to answer the fundamental questions- how the article can advance this field of study? Please also focus on readers' satisfaction and make your research explained in a clear manner not focusing on aspects that are not later presented in the paper.
Author Response
The research presents in this paper a very interesting topic, as well as results that are of wider significance when it comes to the application of vibration measurements in person identification. The paper is, however, not always concise and clear. The literature in the paper is adequately cited.
Authors' Response: We appreciate the reviewer's positive feedback on the topic and the broader significance of the results presented in our research on the application of vibration measurements in person identification. We have taken note of the reviewer's comment regarding the need for improved conciseness and clarity in the paper. To address this concern, we have made significant efforts to enhance the overall clarity and coherence of the manuscript. We have revised the text to ensure that our arguments and findings are presented in a concise and straightforward manner, allowing for easier understanding by the readers. Additionally, we have restructured the paper to improve the flow and logical progression of the content.
We are glad to hear that the literature in the paper is adequately cited. We have made sure to provide proper citations and references to acknowledge the relevant works in the field and support our claims and findings. We appreciate the reviewer's feedback, as it has guided us in refining the clarity and conciseness of our paper. These improvements will contribute to a better understanding of our research and its implications in the field of vibration-based person identification.
In the case of literature, all the cited references are relevant to the research. No unnecessary self-citations were detected.
Authors' Response: We appreciate the reviewer's positive feedback on the literature.
The English language used is without major mistakes, however, some sentences need to be checked by a native speaker.
Authors' Response: We appreciate the reviewer's comments on our paper's English usage. We have made every effort to ensure that the document is free of serious grammatical and linguistic errors. However, we recognize that some sentences may still benefit from the expertise of a native speaker. To address this worry, we had the work read by a native English speaker to identify any parts that still need to be improved. Their knowledge has assisted us in refining the phrasing in those precise sentences, resulting in better correctness and clarity.
The abstract is well written. The role of the abstract is to give a basic overview of the paper. In this case, the abstract gives a good introduction to the paper without specific data and is very informative even for those none-familiar with the topic reader. However, the novelty aspects of the paper are not presented.
Authors' Response: We appreciate the reviewer's positive feedback on the abstract. We have taken note of the comment regarding the novelty aspects of the paper and their absence in the abstract. And therefore, we have revised the abstract to explicitly highlight the novelty aspects of our research. We have emphasized the unique contributions of our study. We are grateful for the reviewer's valuable input, as it has prompted us to enhance the abstract to better convey the novelty and significance of our research.
The reviewer would also suggest not using the abbreviation in the abstract but the full names and introducing the abbreviation the first time it appears in general text.
Authors' Response: We appreciate the reviewer's comment regarding the use of abbreviations in the abstract. We have taken this feedback into consideration and have made the necessary revisions to ensure that all abbreviations used in the abstract are defined and introduced appropriately. Additionally, throughout the text, we have followed the practice of introducing abbreviations by providing their full names upon their first appearance.
It seems that the authors use sometimes impersonal and sometimes personal forms. It is more common to use impersonal forms in scientific papers. Use only one convention in the whole article.
Authors' Response: We appreciate the reviewer's feedback regarding the use of personal and impersonal forms in our paper. We acknowledge that maintaining consistency in the choice of writing style is essential in scientific papers, and we apologize for any inconsistencies observed. We have diligently reviewed the entire manuscript and made the necessary revisions to ensure a uniform writing style throughout. We have carefully converted all instances of personal forms to passive voice or utilized phrases such as "this study" as the subject. We thank the reviewer for pointing out this issue, as it has allowed us to refine the writing style and ensure greater consistency in our paper.
The introduction is mostly well-written and follows all the rules of the proper instruction on the topic. However,
a. In general, the introduction is quite basic and is more typical for a magazine publication than a scientific paper.
b. The last paragraph where the aim, scope and novelty are to be presented is quite short and the novelty aspect is weak or not existing.
Authors' Response: We appreciate the reviewer's comment on the introduction section. We have carefully considered the comments and made substantial revisions to address the raised concerns. To address the first comment, we have expanded the introduction to provide a more comprehensive and in-depth overview of the topic. We have included additional relevant information and contextual details to ensure that the introduction aligns with the standards expected for a scientific paper rather than a magazine publication.
Regarding the second comment, we have dedicated a more substantial paragraph at the end of the introduction to clearly present the aim, scope, and novelty of our research. We have provided a detailed explanation of the unique contributions and innovative aspects of our study to address any perceived weaknesses in this regard, however, the list of the contribution is stated at the end of the results section.
Chapter 2 is well written
Authors' Response: We appreciate the reviewer's positive feedback
Material and methods:
a. no information on the sensor's parameters.
b. No photos of the test conditions would clarify how the test was conducted. No connection between the human orientation with the orientation of the phone. No explanation of how X,Y,Z axis on Fig. 2 are situated during the test. Thus this figure is meaningless.
Authors' Response: We appreciate the reviewer's comments on the "Materials and Methods" section. Regarding the first comment, we have addressed the lack of information on the sensor's parameters by adding a dedicated paragraph in the introduction. This paragraph now provides relevant details about the smartphone sensors used in our study. However, since the data used in our research is obtained from the Hand Movement, Orientation, and Grasp (HMOG) public dataset, we have referred the reader to the original dataset reference for more specific information on the sensor's parameters.
In response to the second comment, we acknowledge the need for photos of the test conditions to provide a clearer understanding of how the experiments were conducted. Unfortunately, as we did not directly conduct the experiments to create the dataset ourselves and relied on the publicly available dataset, we do not have access to the specific photos of the test conditions. However, we have made efforts to provide detailed descriptions of the experimental setup to compensate for the lack of visual aids.
Regarding the issue raised about Figure 2, we appreciate the reviewer's insight and agree that the figure lacks clarity in explaining the orientation of the X, Y, and Z axes during the test. In light of this, we have decided to remove Figure 2 from the paper, as it does not contribute effectively to the presentation of our methodology. We thank the reviewer for bringing these points to our attention and helping us enhance the quality of our research presentation.
Fig 2 and 3 are basic, well know information that brings nothing to the reader except the one unfamiliar with statistical or signal analysis.
Authors' Response: We appreciate the reviewer's feedback regarding Figures 2 and 3. In response to the comment about Figure 2, we agree that it does not provide significant value to readers who are already familiar with statistical or signal analysis. Therefore, we have decided to remove Figure 2 from the revised version of the paper.
Regarding Figure 3, we understand that it may be considered basic and well-known information. However, we believe that it still serves a purpose in illustrating the Used Overlapping Sliding Window during the feature extraction process. While the concept may be trivial, we feel that including this figure can contribute to the reproducibility of our research. It provides clarity on the specific sliding window configuration used in our study, which can vary from one research study to another. By keeping Figure 3 (now Figure 2 in the revised version), we aim to facilitate the replication of our methodology by providing a visual representation of this important aspect. We appreciate the reviewer's perspective on these figures and acknowledge their input. Our intention is to strike a balance between providing useful information to readers and ensuring the reproducibility of our research. By removing Figure 2 and retaining Figure 3, we aim to address the reviewer's concerns while still maintaining the clarity and replicability of our methodology.
The paper does provide proper conclusions. However, again the novelty aspect must be pointed out strongly.
Authors' Response: We appreciate the reviewer's feedback regarding the conclusions section of our paper. To address this concern, we have added a new subsection at the end of the Results section, explicitly highlighting the novel contributions of our research. And another subsection provides a comprehensive overview of the new authentication experiments conducted and the utilization of gyroscope data in conjunction with accelerometer data. These additions significantly enhance the novelty of our work. Additionally, we have rewritten the conclusion section to incorporate these new findings and underscore their significance.
The biggest problem of the paper is that the authors are writing about some solutions like using an accelerometer, gyroscope and magnetometer but after only using an accelerometer. Similarly with the algorithms. This makes the paper not truly thought through and some things difficult to follow.
Authors' Response: Indeed, we acknowledge the reviewer's feedback highlighting the inconsistency in our previous version of the paper. We agree that there was a discrepancy in discussing the potential use of accelerometer, gyroscope, and magnetometer, but ultimately focusing only on the accelerometer in our study. Additionally, the inclusion of algorithms without their subsequent utilization added to the confusion. To address this concern, we have made significant revisions to the paper, we even changed the title of the paper to be “Smartphone User Identification/Authentication Using Accelerometer and Gyroscope Data”. We have added a new subsection that specifically addresses the new authentication experiments using both accelerometer and gyroscope data. This new section provides a detailed comparison between the two sensor types and highlights their respective advantages and limitations. Through this analysis, we have determined that the magnetometer data is influenced by environmental factors and other interferences, which led us to omit its usage in our study. This clarification strengthens the coherence and cohesion of the paper.
B) Item remarks
Some figures are out of focus probably due to pdf compression. Make sure that this is not the case in the final version
Fig .9,10, are hardly readable- please enlarge them and make sure that axis captions and legends are visible.
Response: To enhance the quality of all figures, we have redrawn them using Python and exported them as PDF files with high quality.
Tab. 7 – comment again on the personal/impersonal form. It looks better to write in impersonal form like in tab 7 “ours” looks strange but if changed to “Authors” or “authors solution” looks much more professional.
Response: We have carefully converted all instances of personal forms to passive voice or utilized phrases such as "this study" as the subject, for this particular Table we replaced “ours” by “This study”
C) Conclusions:
The biggest problem of the article is also the clear presentation of the novelty of the research topic and in many places confusing structure or claims that are not followed later in the test or results presentation.
Thus, the reviewer suggests the major corrections mentioned previously and asks the authors to answer the fundamental questions- how the article can advance this field of study? Please also focus on readers' satisfaction and make your research explained in a clear manner not focusing on aspects that are not later presented in the paper.
Authors' Response: We appreciate the reviewer's feedback and recognize the concerns regarding the clarity of novelty and structure in our previous version of the paper. We apologize for any confusion caused. In response to the reviewer's comments, we have made significant revisions to address these issues. We have revised the title of the paper to "Smartphone User Identification/Authentication Using Accelerometer and Gyroscope Data" to better reflect the focus of our research. Additionally, we have added a new subsection that highlights the new authentication experiments conducted using both accelerometer and gyroscope data. And we have provided a comprehensive comparison between these two sensor types, discussing their advantages and limitations. This analysis allowed us to make informed decisions, such as excluding the magnetometer data due to its susceptibility to environmental influences and interferences. By incorporating these changes, we have enhanced the coherence and cohesion of the paper, ensuring that our research is explained in a clear manner and aligns with the claims presented.
We have also taken into consideration the reviewer's emphasis on advancing the field of study. Through our research, we have contributed by conducting extensive experiments on user identification/authentication and demonstrating the effectiveness of combining accelerometer and gyroscope data. This combination has yielded significant improvements in authentication accuracy, which can have practical implications for enhancing security and privacy in smartphone applications.
Additionally, we have focused on readers' satisfaction by revising the structure and presentation of our research, ensuring a clear and logical flow of information. We have addressed the fundamental questions raised by the reviewer and made sure that our claims are supported by the corresponding test results and presented in a coherent manner. We thank the reviewer for their valuable feedback, which has helped us improve the quality and impact of our paper.
Reviewer 3 Report
Please see the document attached to minor correcctions.
Comments for author File: Comments.pdf
It is well writen. But, you should try a based english native to identify unprecetable incorrections.
Author Response
Please see the document attached to minor corrections.
peer-review-29784065.v1.pdf
Response: Thank you for providing the attached document with the suggested corrections. We appreciate your efforts in editing our work, as it has been very helpful in improving the overall quality of the paper. Based on the feedback provided in the attached PDF, we have made the necessary changes. We have removed the words mentioned as "do not need" to streamline the text and improve its clarity. Additionally, we have addressed all the issues with citations as outlined in the document, ensuring proper formatting and accurate referencing.
Furthermore, we have carefully paraphrased the paragraphs that were marked as "incomprehensible," taking into consideration the suggested improvements to enhance the readability and understanding of the content. We have also made the necessary revisions to Tables 1 and 2, implementing the suggested changes to improve their accuracy and presentation.
Once again, we would like to express our gratitude for the thorough review and valuable suggestions provided in the attached document. Your input has significantly contributed to the refinement of our paper.
Comments on the Quality of English Language
It is well writen. But, you should try a based english native to identify unprecetable incorrections.
Response: We appreciate the reviewer's positive feedback regarding the writing, however, we have taken the necessary steps to improve the language quality of the manuscript. A native English speaker has reviewed the article and corrected any identified language issues.
Reviewer 4 Report
The article analysed user identification using smartphone accelerometer data. The authors analyse the opportunity of using the standard technologies using built-in acceleration smartphone sensors to train deep learning and classify methods to recognise the users. The described identification system outperforms existing methods, which were principally created and tested on the same HMOG (Hand Movement, Orientation, and Grasp) public smartphone dataset, even with more users. Additionally, the Authors investigate using some pre-processing methods to clean the data and improve the performance of machine learning.
The bibliography choice linked to the analysed scientific problem is relevant and sufficient. The sued symbols (constants, variables) and functions are described appropriately. However, the graphs should be corrected since they are indecipherable.
The article analysed the crucial problem regarding using sensors built into commonly used devices to solve the scientific issue. However, the scientific analysis soundness of this issue is average. It means the current description of the analysis needs to be revised to gain scientific soundness. The mathematical and technical analysis should be presented more widely and precisely using mathematical matters. Moreover, the Authors should revise the scientific discussion chapter to more widely and precisely analyse the presented outcomes and adequately rephrase the conclusion part.
All in all, the article can be published after a minor revision of the remarks mentioned above.
Author Response
The article analysed user identification using smartphone accelerometer data. The authors analyse the opportunity of using the standard technologies using built-in acceleration smartphone sensors to train deep learning and classify methods to recognise the users. The described identification system outperforms existing methods, which were principally created and tested on the same HMOG (Hand Movement, Orientation, and Grasp) public smartphone dataset, even with more users. Additionally, the Authors investigate using some pre-processing methods to clean the data and improve the performance of machine learning.
Authors’ response: We thank the reviewer for their positive assessment of our article. And we appreciate their recognition of the potential of standard technologies and built-in acceleration smartphone sensors in training deep learning and classification methods for user recognition.
The bibliography choice linked to the analysed scientific problem is relevant and sufficient. The sued symbols (constants, variables) and functions are described appropriately. However, the graphs should be corrected since they are indecipherable.
Authors’ response: We thank the reviewer for their positive feedback. To enhance the quality of the figures, we have redrawn all of them using Python and exported them as PDF files with high quality.
The article analysed the crucial problem regarding using sensors built into commonly used devices to solve the scientific issue. However, the scientific analysis soundness of this issue is average. It means the current description of the analysis needs to be revised to gain scientific soundness. The mathematical and technical analysis should be presented more widely and precisely using mathematical matters. Moreover, the Authors should revise the scientific discussion chapter to more widely and precisely analyse the presented outcomes and adequately rephrase the conclusion part.
Authors’ response: We appreciate the reviewer's feedback on our article. We acknowledge their comment regarding the scientific analysis, and we recognize the need to enhance the scientific soundness of our work. To address this concern, we have conducted a thorough revision of the analysis section. We have expanded the technical analysis, presenting it in a more comprehensive and precise manner. Additionally, we have taken the reviewer's suggestion to heart and revised our scientific discussion. We have provided a wider and more detailed analysis of the presented outcomes, delving deeper into the implications and significance of our findings. We have also carefully rephrased the conclusion part to accurately reflect the key insights and conclusions drawn from our research. We are grateful for the reviewer's valuable feedback, which has prompted us to make substantial improvements to our paper.
All in all, the article can be published after a minor revision of the remarks mentioned above.
Authors’ response: We have thoroughly addressed all the minor revisions and remarks mentioned by the reviewer. Thank you for your valuable feedback, which has greatly contributed to the improvement of our article. With these revisions in place, we believe that the paper is now ready for publication.
Round 2
Reviewer 1 Report
Although the authors have addressed a few of my previous concerns still, the article needs significant improvement.
1. Justify the novelty of the work.
2. Improve the presentation of results.
3. Still figures are poor. There are still Excel graphs that must be replaced with efficient plots. Please replace all Excel plots (Figures 3 to 7 and Figures 11-12)
4. Font size in confusion matrices are not readable. Improve it.
5. Is it compulsory to highlight rows in Table 10?
English language has been improved.
Author Response
Although the authors have addressed a few of my previous concerns still, the article needs significant improvement.
Authors’ response:
Thank you for your feedback. We appreciate the time and effort the reviewer has invested in providing valuable insights to improve our article. We apologize if our response did not fully meet the reviewer's expectations. We have carefully considered and addressed all the concerns and suggestions raised in the previous review round. However, we understand that there may still be areas that require further improvement. We are committed to enhancing the quality of our article, and we would greatly appreciate any specific recommendations or suggestions from the reviewer to help us achieve that goal. We value the reviewer's expertise and insight and will diligently work towards implementing any necessary changes to meet the desired standards. Thank you once again for your time and consideration. We look forward to receiving further guidance from the reviewer to improve our article to its fullest potential.
- Justify the novelty of the work.
Authors response: Thank you for your comment regarding justifying the novelty of our work. We appreciate your feedback and would like to highlight the contributions and novelty of our research.
The novelty and contributions of this work can be summarized as follows:
Investigation of Utility: Our article explores the potential applications and usability of accelerometer data, focusing specifically on user identification and authentication. We aim to provide practical insights into the application of accelerometer data by employing machine learning techniques and employing pre-processing and feature selection approaches.
User Identification: We utilize various machine learning algorithms, including deep learning, standard classifiers, and voting classifiers, to accurately identify users based on accelerometer data. Our analysis and training are performed using the HMOG public dataset.
Data Pre-processing and Feature Selection: We analyze multiple pre-processing approaches and feature selection strategies to enhance the performance of machine learning algorithms. These techniques aim to cleanse and optimize accelerometer data, leading to improved user identification accuracy.
User Authentication: Our study expands the scope beyond user identification to encompass the authentication process. We investigate the use of gyroscope data in conjunction with accelerometer data to explore their combined impact on authentication accuracy.
Sensor Comparison and Analysis: We compare and analyze the effectiveness of accelerometer data alone for authentication and evaluate the influence of gyroscope data on the authentication process. Our comparative analysis highlights the individual and combined contributions of these sensors, demonstrating the superiority of accelerometer data for authentication and the added benefits of combining both sensors. We also omit the magnetometer's data due to its susceptibility to environmental factors.
Overall, our research contributes to a better understanding of the utility of smartphone sensor data, particularly in the fields of user identification and authentication. We provide valuable insights into the potential applications and performance optimization approaches in these domains.
We summarized these at the end of the results section. We hope this clarifies the novelty and contributions of our work. Thank you again for your time and valuable input.
- Improve the presentation of results.
Authors’ response:
Thank you for your comment regarding the presentation of results. We have made improvements to address this concern. Specifically, we have taken the following actions: Improved Figure Resolution: We have enhanced the resolution of all figures presented in the results section. This adjustment ensures that the visual elements are clear and easily interpretable. Enhanced Captions: We have revisited the captions accompanying the figures to provide more detailed and informative descriptions. Removal of Highlight in Comparisons Table: Based on your feedback, we have removed the highlight in the comparisons table. This modification aims to streamline the presentation and avoid any potential confusion caused by excessive highlighting. By implementing these changes, we believe the presentation of results has been significantly improved. We appreciate your valuable input in helping us enhance the clarity and effectiveness of our findings.
- Still figures are poor. There are still Excel graphs that must be replaced with efficient plots. Please replace all Excel plots (Figures 3 to 7 and Figures 11-12)
Authors’ response:
Thank you for your comment regarding the quality of figures in our manuscript. We have taken your feedback into careful consideration and made the necessary improvements. Specifically, we have addressed the issue raised regarding the use of Excel graphs and replaced them with more efficient plots. To ensure the clarity and visual quality of the figures, we have obtained the original images from our Python scripts. We have then increased the resolution of these figures to 1200 dpi. This enhancement guarantees that the details and information conveyed in the figures are presented with higher fidelity and improved readability. We have applied these improvements to all figures mentioned in your comment, including Figures 3 to 7 and Figures 11-12. By replacing the Excel plots with more suitable and visually appealing plots, we have aimed to enhance the overall quality and professionalism of the manuscript. We sincerely appreciate your valuable feedback, which has enabled us to enhance the presentation of our results. Your comments have been instrumental in improving the visual aspects of our manuscript, and we are grateful for your assistance in making our research more accessible and visually engaging.
- Font size in confusion matrices are not readable. Improve it.
Authors’ response:
Thank you for your feedback regarding the font size in the confusion matrices presented in our manuscript. We appreciate your attention to this detail, and we have taken immediate action to address this concern. To ensure readability and clarity, we have increased the font size in the confusion matrices. This adjustment allows readers to easily discern the values and make accurate interpretations of the results. Additionally, we have also increased the size of the confusion matrices themselves, further improving the legibility of the presented information. Furthermore, we have paid close attention to the resolution of the confusion matrices to guarantee optimal visibility. By enhancing both the font size and resolution, we aim to provide a better reading experience and facilitate a comprehensive understanding of the classification performance. We are grateful for your insightful suggestion, as it has led to the improvement of the visual presentation of our results. Your attention to detail and commitment to enhancing the quality of our manuscript are genuinely appreciated.
- Is it compulsory to highlight rows in Table 10?
Authors’ response:
Thank you for your comment regarding the highlighting of rows in Table 10. We appreciate your feedback and would like to address this concern. Highlighting the rows in Table 10 was not compulsory. However, in the first round of reviews, the addition of these rows was requested by one of the reviewers to provide further clarity and context. To distinguish these added rows from the original content, we chose to highlight them. We understand that this highlighting may not be essential and could potentially distract from the overall presentation. Therefore, based on your suggestion, we have revised the table and removed the highlighting from the added rows. This modification ensures consistency throughout the table and maintains a clean and professional appearance. We appreciate your attention to detail and your commitment to improving the quality of our manuscript. Your feedback has been invaluable in refining our work.
Reviewer 2 Report
Dear authors,
The manuscript has been extensively corrected in accordance with most of the comments from the reviewers.
Some small elements still can be improved:
• The affiliation list seems not unified in the case of the information presented. Please unify according to the template of the journal
• The abstract Is good. No further changes are necessary
• The introduction was improved. However, lines 20-45 present general well-known usage of mobile phones. Would suggest focusing more on the engineering usage of mobile phones and their sensors. Additionally, some examples of sensor usage are presented in case of accelerometer, magnetometer, and gyroscope, however, the authors are missing one very important sensor – the microphone which can be used, as recent studies show, e.g. for quality control purposes (e.g. DOI: 10.3390/s23125639) or for noise/sound lever measurements (e.g https://doi.org/10.1016/j.apacoust.2018.04.011). Please improve and focus mostly on engineering applications as your article is also presenting a unique engineering application.
• Most of the other chapters were improved. However, in Fig. 10 the axis captions and legends are not visible and can not be read (must be improved). Would also suggest improving text readability for Fig. 13
• Conclusions were corrected and the reviewer has no additional requests.
Hope this will solve all, still existing issues.
Best regards,
The reviewer.
Author Response
The manuscript has been extensively corrected in accordance with most of the comments from the reviewers.
Some small elements still can be improved:
• The affiliation list seems not unified in the case of the information presented. Please unify according to the template of the journal
The abstract Is good. No further changes are necessary
• The introduction was improved. However, lines 20-45 present general well-known usage of mobile phones. Would suggest focusing more on the engineering usage of mobile phones and their sensors. Additionally, some examples of sensor usage are presented in case of accelerometer, magnetometer, and gyroscope, however, the authors are missing one very important sensor – the microphone which can be used, as recent studies show, e.g. for quality control purposes (e.g. DOI: 10.3390/s23125639) or for noise/sound lever measurements (e.g https://doi.org/10.1016/j.apacoust.2018.04.011). Please improve and focus mostly on engineering applications as your article is also presenting a unique engineering application.
Most of the other chapters were improved. However, in Fig. 10 the axis captions and legends are not visible and can not be read (must be improved). Would also suggest improving text readability for Fig. 13- Conclusions were corrected and the reviewer has no additional requests.
Authors' response:
Thank you for your feedback on our revised manuscript. We appreciate your thorough evaluation and suggestions for further improvement. We have carefully addressed the remaining issues you raised, and we would like to provide our responses below:
Affiliation list: We have unified the affiliation list according to the template of the journal, ensuring consistency and accuracy.
Abstract: We are glad to hear that the abstract is in good shape and requires no further changes.
Introduction: We have taken your suggestion into account and focused more on the engineering applications of mobile phones and their sensors. Additionally, we have included examples of sensor usage, including the microphone, as highlighted in recent studies such as the ones you mentioned (DOI: 10.3390/s23125639 and https://doi.org/10.1016/j.apacoust.2018.04.011). These additions enhance the engineering perspective of our article.
Fig. 10 and Fig. 13: We have improved the visibility of the axis captions and legends in Fig. 10 to ensure readability. Furthermore, we have enhanced the text readability for Fig. 13, making it more accessible to readers.
Conclusions: We are pleased that you found the conclusions to be in order with no additional requests.
We believe that these revisions have significantly improved the manuscript, aligning it more closely with the objectives of the journal and addressing your valuable feedback. Thank you again for your time and thorough evaluation.
Round 3
Reviewer 1 Report
The authors have addressed my previous comments however in the revised version all the figures have inconsistent font size. It must be improved before publication.
Moderate editing of English language is required.
Author Response
Thank you for your comment, this happened because of the increase in the resolution of the images.
could you please leave this to the publication production team?