Enhancing Fingerprint Forensics: A Comprehensive Study of Gender Classification Based on Advanced Data-Centric AI Approaches and Multi-Database Analysis
and Shimon Kimchi 3Round 1
Reviewer 1 Report
Comments and Suggestions for Authorsa) The Abstract is too long. The first half of the Abstract justifies the work carried out and this text makes sense to be in the Introduction chapter and not in the Abstract.
b) In this work, the authors propose to identify the type of sex using fingerprints achieved in a crime. The authors do not justify the relevance of this information for solving the crime, since even if an algorithm presents that there is a 99% probability of the fingerprint belonging a woman, it does not guarantee 100% that it is from a woman, leading to the crime investigators have to consider the possibilities of being a man or a woman. In other words, while identifying the type of sex may be scientifically relevant, it does not appear to be relevant in a crime when the probability is less than 100%. It is recommended that authors review the motivation written in the Abstract and in chapter 1-Introduction.
c) The main contributions described at the end of chapter 1-Introduction are very generic and contain statements frequently presented in works. In the main contributions, it is expected to describe what is new in this work and which has not been presented in other works; in other words, what is innovation.
d) Authors must be concerned with the presentation of the document/work and avoid situations such as having a table (in this case, table 1), split between two pages. They should also review the formatting applied in the References chapter, which is 2.0 lines or 1.5 lines spacing and not 1.0 lines. Still on the references, they must correct the information presented in the references, as [2] [6] [8] (…) it is not clear whether it corresponds to a publication in an international Journal or a Conference or a Book or it is a Web page. All references must be checked to avoid situations such as [8] whose third author is “A. Professor” and this third author does not exist, since the original document has only two authors. Text such as “Top of Form” and “Bottom of Form” that appears at the beginning of the Methods chapter must be removed. These errors suggest that the authors did not carefully read the document before submitting.
e) The methodology uses the central region of the fingerprint and the region of the ring that surrounds the central region. Authors must explain the criteria used to define these regions, especially when images with different resolutions are used. They should also mention the behavior of region selection when using images achieved at a crime scene, in which the fingerprints are not aligned (and have a rotation with a random angle), or when the images are off-center (the central region is in a corner of the image), or even when the image only has 50% of the fingerprint (there is only half of the central region).
f) The Discussion chapter actually corresponds to conclusions, as the authors draw conclusions from the results. Given that the Conclusions chapter is very small and repeats the information from the Discussion chapter, I recommend that the Discussion + Conclusions chapters be merged, producing a new chapter called Conclusions. Note that it is not just joining the text, it is joining and re-reading the text to remove repeated statements.
Author Response
Please see the attachment."
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe article is interesting and well-prepared. However, there are some remarks that should be implemented in the revised version of the article:
1. Some minor editorial remarks:
- On page 2, the paragraph starting with 'between populations' is redundant; consider merging it with the previous one.
- On page 2, there is a missing dot after the words 'favorable outcome.'
- On page 3, a dot is missing after the section title "Datasets."
- On page 4, the phrases 'Top of Form' and 'Bottom of Form' seem out of context and may need clarification or correction.
- Figures 2 and 3 could benefit from higher resolution.
2. The sentence 'To train our CNN models we employ standard training tech-
niques [11], including data augmentation, normalization, a method for class imbalance [24], stochastic AdaGrad optimization and others.' requires more detailed explanation. What kind of augmentation was used? Which method for improving class imbalance was implemented? Was cross-validation employed? What were the sizes of the train/test/validation datasets after augmentation?
3. The literature review could be more detailed. Consider including more recent references, as the article currently includes only one position from 2023.
Author Response
Please see the attachment."
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors made changes to the work in accordance with this reviewer's comments. Consequently, the overall quality of work has increased considerably.
