A Systematic Literature Review on Artificial Intelligence and Explainable Artificial Intelligence for Visual Quality Assurance in Manufacturing

Round 1

Reviewer 1 Report (Previous Reviewer 3)

Authors improved their work and put a great effort into writing this systematic review. Thank you.   

Author Response

Dear Sir or Madam,

we are glad that you are now satisfied with our work and thank you for your help to improve our quality. After the additional effort we put into this work, the paper is much more comprehensive and provides more insights into this field.

Kind regards

Prof. Dr. Christoph Reich

Rudolf Hoffmann 

Reviewer 2 Report (Previous Reviewer 2)

Thanks for the revision and re-submission. I would like to follow up with the following two comments:
It lacks discussion on benchmark datasets or experimental pipelines.
It lacks a discussion on performance metrics or experimental comparisons between various references.
I would like to suggest adding a section to explicitly discuss these two points.

NA

Author Response

Dear Reviewer,

thank you for giving us the opportunity to submit a revised draft of my manuscript titled A Systematic Literature Review on Artificial Intelligence and Explainable Artificial Intelligence for Visual Quality Assurance in Manufacturing to Electronics. We appreciate the time and effort that you have dedicated to providing your valuable feedback on our manuscript. We are grateful to you for your insightful comments on our paper. We have been able to incorporate changes to reflect most of the suggestions that you have provided. We have highlighted the changes within the manuscript with the color cyan.

Here is a point-by-point response to your comments and concerns.

Comment 1: It lacks discussion on benchmark datasets or experimental pipelines.

Response 1: Thanks for this remark. Using your comment, we were able to enrich our results. We included an additional section to present the benchmark datasets (see section 6.4 or results4.tex). We added this section as subsection of the results and presented, for what VQA practices these datasets were used.

Comment 2: It lacks a discussion on performance metrics or experimental comparisons between various references.

Response 2: We added a subsection to present the basics and the evaluation focus of the performance metrics that were used in the resulted literature. We added the subsection “Performance Metrics for VQA” to the section “Machine Learning” (see section 2.2 or ML.tex line 34ff).

Sincerely

Prof. Dr. Christoph Reich,

Rudolf Hoffmann

Reviewer 3 Report (Previous Reviewer 1)

The authors have successfully addressed all the comments raised by the reviewers. No further comments

Good

Author Response

Dear Sir or Madam,

we are glad that you are now satisfied with our work and thank you for your help to improve our quality. After the additional effort we put into this work, the paper is much more comprehensive and provides more insights into this field.

Kind regards

Prof. Dr. Christoph Reich

Rudolf Hoffmann 

Round 2

Reviewer 2 Report (Previous Reviewer 2)

The revision is fine. I don't have further comments.

NA

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

the paper is interesting and in an important topic in the manufacturing domain. There are several issues that need to be addressed prior to acceptance. Please find below some comments to improve the quality of the paper:

1) why authors did not search on scopus as one of the largest scientific databases which includes also science direct? also what about web of science, also very aligned with the topic. It is suggested that authors search also those databases and update the results.

2) Since quality has a significant role in the paper. It is suggested that authors place their work into a broader concept, such as Zero Defect Manufacturing (ZDM).  It is suggested to use the following references. Positioning the paper to a higher framework authors will make their work more complete and increase significantly the impact of their paper. ZDM is composed out of four strategies “Detect” (physical or virtual), “Predict”, “Repair” and “Prevent”. Below are suggested some key research works in ZDM domain. Also it is suggested not to limit on the suggested ones, those are suggested as the foundation of ZDM.

a.     Psarommatis, F., May, G., Dreyfus, P.A. and Kiritsis, D., 2020. Zero defect manufacturing: state-of-the-art review, shortcomings and future directions in research. International Journal of Production Research, 58(1), pp.1-17. https://doi.org/10.1080/00207543.2019.1605228

b.       Psarommatis, F., Sousa, J., Mendonça, J.P. and Kiritsis, D., 2021. Zero-defect manufacturing the approach for higher manufacturing sustainability in the era of industry 4.0: a position paper. International Journal of Production Research, pp.1-19. https://doi.org/10.1080/00207543.2021.1987551

3)The paper lack scientifc justification. Why this paper is important and novel, authors do not explain it in details. 

4) Authors should also look for similar review papers that analysis the same domain, present briefly their scope, results, and directions. This is needed for identifying the reaserch gap which the current paper is filling

5)the paper looks like a bibliometric analysis rather than a literature review, meaning that the only results are based on statistics for the collected sample papers. The results needs to be enriched with more insights from the papers.

6) sections 2 and 3 is not clear whether they are an outcome from the current paper or comming from the literature.

7) the discussion could benefit from enrichment with more discussion points and challenges in the domain

8) it is suggested that authors aslo search for the "virtual metrology" term as is one of the leading inspection methods using AI. There is a recent literaure review paper on this topic that you can use.

Minor improvements are needed

Reviewer 2 Report

This paper presents a review of AI and XAI in visual quality assurance in manufacturing. Both are big topics. The major findings are presented in Table 3.
It is just a quantitative review by simply listing out the number of papers in different categories. There are many qualitative analyses missing. For example,
- There is no comparison between this review and the existing reviews (if any) on the same topic.
- There is no link between the discussed papers and the three definitions presented in Section 2.
- the various XAI techniques discussed in Figure 3, there is no discussion on what the unique challenges of visual quality assurance in manufacturing are or how to apply XAI to address these challenges.
- There is no discussion on benchmark datasets or experimental pipelines.
- There is no discussion on performance metrics or experimental comparisons between various references.

NA

Reviewer 3 Report

Remark 1:

Researchers made great effort with this research paper. Its comprehensive and very detailed study, BUT I do not think they understand so well the foundations of Artificial Intelligence and the Machine Learning.

Dear Authors, all these models relay on a complex mathematical foundation and if possible at least some general information about their foundation and how do they operate, how do they distinguish from each other, why some of them are used for classification and the others for other can be provided | Please provide it, where most suitable in the paper.

Just for instance, logistic regression is used for binary classification, SVM maps not linearly separable problems into higher dimension by using some type of kernels where the problem becomes separable, … etc.

Remark 2:

Section: Artificial Intelligence Facilitates Visual Quality Assurance

Sentence: By leveraging techniques such as ML (citations needed? please provide),

When mentioned ML, please provide some references that apply it to solve some real problems in different domains, such as:

1.      For handwriting recognition:

10.1109/ICFHR.2012.179

2.      In medicine:

https://doi.org/10.1016/j.jksus.2023.102573

3.      In agriculture:

….