Personal Identification Using Embedded Raspberry Pi-Based Face Recognition Systems

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 

This paper proposed a personal identification method using embedded Raspberry Pi-based face recognition systems. The following suggestions are given for this paper：

1. The fifth and sixth pictures in this article are not clear enough. Enhancing the resolution or selecting more defined images would greatly benefit the understanding of the content.

2. There is no subscript for y_ih (x) in line 296, and there is a problem with the punctuation of y·h(x) in line 298. Ensuring proper subscript usage and correcting punctuation will aid in maintaining clarity and professionalism in the text.

3. AdaBoost machine learning lacks the final weighted voting part. This is a fundamental aspect of the AdaBoost method, which combines multiple weak classifiers to form a strong classifier by assigning weights to each classifier based on its performance. 

4. No punctuation on line 307. Please check and correct the punctuation marks of the full text carefully.

5. Why should references be used in figures 6 and 7 and what is the role? Please clarify why references are used in these specific figures and explaining their significance.

6. The content on the Cascading Haar classifier lacks sufficient mathematical formulas to fully explain its workings. Too few formulas for Cascading Haar classifier.

7. Your system is based on the Raspberry Pi, but I find little raspberry pi related techniques in your article. You need add some content about the display of face recognition on Raspberry Pi device.

Author Response

Dear Associate Editor and Reviewers,

The authors would like to thank the reviewers and the associate editor for their effort in reviewing the manuscript and for their valuable and constructive comments and fruitful observations, which helped in improving the quality of the manuscript to a publishable standard. In attached files are the responses to the reviewers’ comments and suggestions. Reviewers’ questions and comments are in BLACK and the authors’ answers and comments are in RED.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This paper demonstrates the potential of a Raspberry Pi-based facial recognition system in the fields of security control and surveillance, and it has a certain degree of technical innovation. However, there is room for further improvement in some aspects of detail expression, which could enhance the quality of the paper.

 

Abstract

(1)     It is suggested to briefly mention the experimental conditions in the abstract, and to clarify under which environments the recognition time of "10.5 seconds" was obtained. Additionally, if there are other key data (such as accuracy, false recognition rate), they should also be mentioned in the abstract to outline the main contributions and experimental outcomes of the paper.

 

1 Introduction

(1)     In the Introduction section, existing research challenges in Privacy concerns, Algorithmic bias and discrimination, and variable conditions are discussed, but it does not explain how this study addresses these issues.

(2)     It is suggested that the Introduction should more prominently highlight the necessity and significance of this study, provide a more detailed explanation of the differences between this research and others, as well as its strengths and weaknesses.

 

 

3 Verification System

 

(1)     In the section on selecting the similarity threshold in 3.1, it is mentioned, "To determine the optimal threshold, several tests were conducted using a group of 398 six male individuals with varying degrees of biometric similarity." It is recommended to elaborate on the differences in these characteristics and the reasons for selecting these individuals.

(2)     In the section 3.2 Effectiveness of recognition depending on facial expressions, is the range of expressions used for recognition comprehensive enough? There is also a lack of specific descriptions for each image's expression, and it is suggested that descriptions be added to avoid misunderstandings among readers.

(3)     In the section 3.3 'Performance and Effectiveness Test of the Face Recognition System', the subjects used for testing are concentrated around the age range of 15-50, with a lack of testing for individuals under 15 and over 50 years old. It is recommended to explain the reasons for the absence of subjects in these age groups.

Author Response

Dear Associate Editor and Reviewers,

The authors would like to thank the reviewers and the associate editor for their effort in reviewing the manuscript and for their valuable and constructive comments and fruitful observations, which helped in improving the quality of the manuscript to a publishable standard. In attached files are the responses to the reviewers’ comments and suggestions. Reviewers’ questions and comments are in BLACK and the authors’ answers and comments are in RED.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Some points which are described below are recommended to be checked  and improved in a revised version: 

-       Eqn.(1): The way (or ways) of defining the functions h/subt(x) has to be explained more clearly.

-       The upper- term of the sum “T” in (1) is not explained.

-       Eqn. (2) and Fig. 6: Alpha has +/-   -infinity asymptotes for the error epsilon values  =0, 1. It is not clearly explained how is this fact addressed when applying the formula.

-       Eqn. (3): What is the argument (i)?- Please, explain it more clearly. Also, give more details about  how it  is evaluated y/subt and which is the concept behind this argument.

-       It is not very clear the paragraph of text after Eqn. (3). For instance, is it meant that y/t  has the same value (+ 1 or -1) as h/t(x)?. Also, if H(x) is + 1 or -1 since it is a sign ( but computed after evaluating a sum of terms) , why (y/t)*(h/t(x)) is also a sign or at least it seems to have the value of a sign ?.

-       Does the weight vector estimates converge from updating (3) or not?,  and. also,  which are the properties of such an updating scheme related to boundedness, convergence, etc.?.

-       The section of conclusions is written in a very schematic and readable way. However some of the commented items are not, in our opinion, very clearly  described in the main body as, for instance, the hardware impact.

-       The introduction  section should describe in more detail the organization and content of the rest of the paper.

-       The organization, grammar and syntax of the material are very acceptable.

 

Author Response

Dear Associate Editor and Reviewers,

The authors would like to thank the reviewers and the associate editor for their effort in reviewing the manuscript and for their valuable and constructive comments and fruitful observations, which helped in improving the quality of the manuscript to a publishable standard. In attached files are the responses to the reviewers’ comments and suggestions. Reviewers’ questions and comments are in BLACK and the authors’ answers and comments are in RED.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed all my comments.

Author Response

Thank you for your kind feedback and for taking the time to review our manuscript. We appreciate your efforts and are glad that our revisions have addressed your comments satisfactorily.

Reviewer 2 Report

Comments and Suggestions for Authors

Abstract

(1)     The content added by the authors in the abstract mainly describes how the accuracy and efficiency of the system perform under different environmental conditions, and how the performance of the system can be enhanced through hardware upgrades and algorithm improvements. However, it is not clear in what specific environment the recognition time of "10.5 seconds" was obtained, nor does it mention key data such as accuracy and false recognition rate. Therefore, this paragraph does not directly answer my question.

 

1 Introduction

(2)     In the added content of the introduction, the author discusses the challenges of the existing research, puts forward a clear solution, and highlights the necessity, significance, differences, advantages and disadvantages of this research, which reasonably answers my questions.

 

3 Verification System

(1)     While the authors detail the key biometric changes considered in selecting the sample group in their response, they mention that a total of 12 participants participated in the study, including eight males of various ages (16 to 49 years) and four females of various ages (19 to 51 years). All the users are recorded in the database. However, in the description of the subsequent results, there is no reference to women. Therefore, the author does not answer my question directly. It is recommended to add this part of the results.

(2)     In the reply, the author described in detail the process of facial expression capture for each member of the sample group. These descriptions provided specific information about the range of expressions and the process of image capture, increased the transparency and understanding of the research method, and reasonably answered my questions.

(3)     The authors did not cite a specific reason for the lack of individuals younger than 15 and older than 50 in the test. Therefore, the author does not answer my question directly.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

The authors would like to thank the reviewer and the associate editor for their effort in reviewing the manuscript and for their valuable and constructive comments and fruitful observations, which helped improve the manuscript's quality to a publishable standard. Detailed below are the responses to the reviewers’ comments and suggestions.

Author Response File: Author Response.docx