A Systematic Review of Reimagining Fashion and Textiles Sustainability with AI: A Circular Economy Approach

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This review paper is quite interesting. It reviews the application of AI and machine vision in textile production and its quality. It also addresses the challenges and limitations of using AI in such applications. It is worthwhile for readers. However, I suggest the following queries should be clarified before accepting for publication.

a. Address the relationship between AI and computer vision technology for textile sorting process.

b. Missing some words in the figure 1.

c. Remove Figure 3 which is not related to the study objectives.

d. Address the application of Generative AI such as ChatGPT and DeepSeek in textile recycling and reuse in the section of "Future Direction".

Comments for author File: Comments.pdf

Author Response

 We sincerely thank the reviewer for their thoughtful feedback and valuable suggestions. Your insights have helped us improve the clarity, quality, and overall impact of our work. We appreciate the time and effort you dedicated to reviewing our manuscript.

1. Missing some words in the figure 1.

Response: The figure has been revised for greater detail, aligning with PRISMA checklist requirements. 

1. Remove Figure 3 which is not related to the study objectives.

Response: 

Including this graph is important because it provides insight into the geographical distribution of research efforts related to textile recycling and the circular economy. Highlighting China's dominant contribution underscores its leading role in this area, which may reflect national priorities, investment in sustainability, or the scale of its textile industry. Understanding which countries are contributing most to the research can help identify global trends, potential collaboration opportunities, and gaps where further research or policy focus is needed. It also contextualizes the global engagement with circular economy principles in textile recycling. 

1. Address the application of Generative AI such as ChatGPT and DeepSeek in textile recycling and reuse in the section of "Future Direction".

Response: The following text has been added in the future work section about generative AI.
In addition, generative AI models, such as Generative Adversarial Networks (GANs) and diffusion models, can be leveraged to synthesize realistic images of garments under various conditions. These generated images can simulate different types of wear, damage, or aging, providing valuable training data for AI systems tasked with assessing garment quality or predicting durability. This approach not only augments existing datasets but also helps in developing more accurate and resilient models for evaluating the condition of textiles in recycling and reuse contexts. 

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a well-organized and timely systematic review of the applications of Artificial Intelligence (AI), particularly computer vision and machine learning, in addressing sustainability issues in the fashion and textiles (F&T) industry. The authors focus on how AI can support a transition to a circular economy (CE), particularly in the areas of garment defect detection, textile sorting, and second-hand garment reuse. The article is well-written and offers valuable insights. But there are several key points where the manuscript must be strengthened to improve its clarity and practical value.

1. I noticed that considering the various AI applications, it does not offer a clear pathway for how these technologies can be implemented in real-world textile reuse or recycling environments.
2. The PRISMA diagram (Figure 1) could be more clearly labeled with numbers of papers at each stage of inclusion/exclusion.
3. Some abbreviations (e.g., CE, CV, DL) appear before they are defined. The whole manuscript needs to be clarified.
4. The graphical breakdown of dataset types (Figure 4) is helpful, but might benefit from a more detailed legend or explanatory note.
5. There is some redundancy in describing dataset limitations across sections; consider consolidating in one focused discussion.
6. The introduction section is poor and needs to includes these references “Impact of sustainable practices on sustainable performance: The moderating role of supply chain visibility” and “Mediating role of eWOM's in green behavior interaction and corporate social responsibility: a stakeholder theory perspective”, and “Green Manufacturing for a Green Environment from Manufacturing Sector in Guangdong Province: Mediating Role of Sustainable Operations and Operational Transparency”
7. Lastly, this article offers a significant contribution to the field of sustainable textiles by synthesizing the current landscape of AI applications in garment quality assessment and textile sorting. However, to elevate the review's practical and scholarly impact, I recommend minor to moderate revision focusing on the points outlined above.

Comments on the Quality of English Language

This manuscript presents a well-organized and timely systematic review of the applications of Artificial Intelligence (AI), particularly computer vision and machine learning, in addressing sustainability issues in the fashion and textiles (F&T) industry. The authors focus on how AI can support a transition to a circular economy (CE), particularly in the areas of garment defect detection, textile sorting, and second-hand garment reuse. The article is well-written and offers valuable insights. But there are several key points where the manuscript must be strengthened to improve its clarity and practical value.

1. I noticed that considering the various AI applications, it does not offer a clear pathway for how these technologies can be implemented in real-world textile reuse or recycling environments.
2. The PRISMA diagram (Figure 1) could be more clearly labeled with numbers of papers at each stage of inclusion/exclusion.
3. Some abbreviations (e.g., CE, CV, DL) appear before they are defined. The whole manuscript needs to be clarified.
4. The graphical breakdown of dataset types (Figure 4) is helpful, but might benefit from a more detailed legend or explanatory note.
5. There is some redundancy in describing dataset limitations across sections; consider consolidating in one focused discussion.
6. The introduction section is poor and needs to includes these references “Impact of sustainable practices on sustainable performance: The moderating role of supply chain visibility” and “Mediating role of eWOM's in green behavior interaction and corporate social responsibility: a stakeholder theory perspective”, and “Green Manufacturing for a Green Environment from Manufacturing Sector in Guangdong Province: Mediating Role of Sustainable Operations and Operational Transparency”
7. Lastly, this article offers a significant contribution to the field of sustainable textiles by synthesizing the current landscape of AI applications in garment quality assessment and textile sorting. However, to elevate the review's practical and scholarly impact, I recommend minor to moderate revision focusing on the points outlined above.

Author Response

We sincerely thank the reviewer for their thoughtful feedback and valuable suggestions. Your insights have helped us improve the clarity, quality, and overall impact of our work. We appreciate the time and effort you dedicated to reviewing our manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. Comment:

1.     I noticed that considering the various AI applications, it does not offer a clear pathway for how these technologies can be implemented in real-world textile reuse or recycling environments.  Response:

The existing limitation of studies examined in this literature reviews identified there were gaps regarding how these technologies have been studied with ramifications for how they can be implemented.

2.     The PRISMA diagram (Figure 1) could be more clearly labeled with numbers of papers at each stage of inclusion/exclusion.

Response: The figure has been revised in greater detail, aligning with PRISMA checklist requirements. Please check figure 1 in the manuscript.

3.     Some abbreviations (e.g., CE, CV, DL) appear before they are defined. The whole manuscript needs to be clarified.

Response: Thank you for pointing out that. We have ensured that all abbreviations defined before their appearance, for example, Circular Economy (CE) in line 60, Computer Vision (CV) in line 72, and Deep Learning (DL) in line 73. Additionally, we have made sure that all abbreviations are clearly and consistently presented throughout the manuscript. List of abbreviations is lso available before references.

4.     The graphical breakdown of dataset types (Figure 4) is helpful, but might benefit from a more detailed legend or explanatory note.

Response: Thank you once again for your valuable feedback and for helping us enhance the quality of the article. The following caption has been added to Figure 5 for more clarity.  We have added another figure to the article. That’s why this figure is numbered 5.

This Figure illustrates the various types of textile samples examined across studies, with fabric swatches being the most commonly used (n=20), followed by full garments, fabric sections, and garment details. Less frequently used samples include trims, microscopic fabric, draped fabric, and yarn. These samples were analyzed using a variety of imaging techniques, including microscopic and hyperspectral imaging, as well as whole garment photography.

5.     There is some redundancy in describing dataset limitations across sections; consider consolidating in one focused discussion.  Response: Thank you for the suggestion. While we acknowledge that dataset limitations are discussed in multiple sections, this was done intentionally and in a consistent manner to highlight how these limitations manifest across different aspects of AI applications. This approach helps reinforce the significance of the issue and ensures that the challenges surrounding datasets become increasingly clear as the discussion progresses. However, we will revisit the structure to ensure clarity and reduce any perceived redundancy. 6.     The introduction section is poor and needs to includes these references “Impact of sustainable practices on sustainable performance: The moderating role of supply chain visibility” and “Mediating role of eWOM's in green behavior interaction and corporate social responsibility: a stakeholder theory perspective”, and “Green Manufacturing for a Green Environment from Manufacturing Sector in Guangdong Province: Mediating Role of Sustainable Operations and Operational Transparency”  Response: Thank you for your thoughtful suggestions. While the papers you mentioned fall slightly outside the specific scope of this study—particularly in terms of the intersection between fashion and textiles, AI applications, and the Circular Economy—we recognize their relevance and value. We believe these works are well aligned with the direction of our future research, and we will be pleased to consider and include them in our upcoming publications where appropriate.    Lastly, this article offers a significant contribution to the field of sustainable textiles by synthesizing the current landscape of AI applications in garment quality assessment and textile sorting. However, to elevate the review's practical and scholarly impact, I recommend minor to moderate revision focusing on the points outlined above.  Response: Thank you once again for valuable suggestions. A revise version has been submitted.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In principle, it is not very clear what the authors wanted to say with the article, what is the purpose, novelty and relevance of the article. The content of the article is not very clear.
If the quality of materials is analyzed, then the analysis is definitely too small and incomplete.
When analyzing the quality of production - the quality of seams, the analysis is also too small and undeveloped.
The conclusions of the article are short and incomplete.
In Figure 1, the end of the text is covered in some boxes, the idea presented in the box is unclear.

Author Response

In principle, it is not very clear what the authors wanted to say with the article, what is the purpose, novelty and relevance of the article. The content of the article is not very clear.
If the quality of materials is analyzed, then the analysis is definitely too small and incomplete.
When analyzing the quality of production - the quality of seams, the analysis is also too small and undeveloped.
The conclusions of the article are short and incomplete.
In Figure 1, the end of the text is covered in some boxes, the idea presented in the box is unclear.

Response:

Thank you for your insightful feedback and for taking the time to thoroughly review our manuscript. We truly appreciate your comments, as they provide us with valuable opportunities to improve our work. 

Regarding your comment about the clarity of the article's purpose, novelty, and relevance, we acknowledge that the interdisciplinary nature and specific focus on the fashion and textiles industry may not be immediately apparent to those outside of this field. Our study is novel and important in its specific application of AI and machine vision to address the unique and pressing challenges within this sector, particularly in minimizing textile waste and supporting circular economy initiatives. This focus is crucial because the fashion and textiles industry is a significant contributor to environmental pollution, and AI offers powerful tools to mitigate its impact. We have revised the introduction to more clearly articulate these points and provide additional context on the significance of this focus. 

We understand your concerns about the scope of the material and production quality analyses.  As a systematic literature review, our analysis is dependent on the breadth and depth of the existing research.  We acknowledge that the number of studies focusing on specific aspects, such as seam quality, is limited, and we have explicitly highlighted this as a gap in the literature and an area for future research.  This also underscores the novelty of our review in identifying these specific gaps. 

Finally, we apologize for the issue with Figure 1.  We have corrected the figure to ensure that all text is clearly visible and that it accurately reflects the PRISMA guidelines.    

We believe that these revisions have significantly strengthened the manuscript, and we are grateful for the opportunity to address your concerns. Thank you again for your time and expertise. 

Reviewer 4 Report

Comments and Suggestions for Authors

This systematic review explores the applications of AI in evaluating clothing quality within the framework of a circular economy, with a focus on supporting second-hand clothing resale, charitable donations by NGOs, and sustainable recycling practices. By analyzing the effectiveness and challenges of related peer-reviewed articles, conference papers, and technical reports, this study highlights state-of-the-art methodologies such as convolutional neural networks (CNNs), hybrid models, and other machine vision systems. A critical aspect of this review is the examination of datasets used for model development, categorized and detailed in a comprehensive table to guide future research. The findings emphasize the potential of AI to enhance quality assurance in second-hand clothing markets, streamline textile sorting for donations and recycling, and reduce waste in the fashion industry. This review concludes with insights into future research directions and the promising use of AI within fashion and textiles to facilitate a transition to a circular economy.

 

The methodology used is sound for a systematic review and is well explained.

The Introduction provide a good framework supported by sufficient number of relevant literature, discusses the potential of AI in sustainability and circular economy in textile and fashion sector and concludes with a well detailed research question. 

The results of the 49 scientific papers retained (out of 135 found) are critically discussed highlighting their shortcoming and strengths. The dataset availability, the applicability and the change of being adopted by SMEs and NGOs due to the complexity of the AI technique used, societal and economic impact are just few criteria’s used to critically discuss the results of the selected studies.

The review further highlights challenges and limitations of AI in textile and fashion industry, circular economy in particular, provides a breakdown of the studies per type of textile samples and country, and provides future research directions.  

The paper is very well structured and written in very good English.

I can suggest some small amendments in layout and typo’s to be addressed:

• Table 2, 3^rd column: harmonization is need, that all sentences start with capital letters;
• Table 3: similar to remark above; some lines of the tables are thicker than the others; if not on purpose, it should be amended
• Some spaces between the words to be deleted on: line 304, 322, 424, 489, 533
• Title of section 3.1.2 could perhaps be slightly changes, as one study is indeed about stich detection but the other it’s about stains (unless the authors means stains occurring during stitching?)

Perhaps the authors could investigate in a future systematic review the potential of AI in assessing garment comfort (both thermophysiological and sensorial). This is a very relevant aspect of clothing, especially in case of high protection clothing (PPEs), but the comfort assessment methods are cumbersome, time consuming, complex and not accessible to clothing sector, dominated by SMEs. Different AI-techniques are potentially useful in predicting garment comfort.

In my opinion, the number of such studies is quite limited, but the authors have experience in systematic reviews and could therefore identify sufficient studies to highlight the strength and the weaknesses of this AI-techniques in this topic. There are many research groups invegating comfort and such systematic review could be relevant for them.

Author Response

We sincerely thank the reviewer for their thoughtful feedback and valuable suggestions. Your insights have helped us improve the clarity, quality, and overall impact of our work. We appreciate the time and effort you dedicated to reviewing our manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.
Kindly find the attached file for detail response.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have made good effort to revise the whole article but still has some issues with the systemic review process. For example,

Point 1, I think still the paper has major problem for data set. This study is restricted to the diversity of datasets which utilized in the studies reviewed. Many datasets were narrowly focused on specific textile types or garment conditions, which diminishes the generalizability of the AI models across a broader range of textiles and real-world scenarios. Please provide a justification about this?

Secondly, the included data set for 49 studies, from these studies, numerous studies failed to provide detailed information on fabric types, structures, or fiber compositions, which are essential for understanding the applicability of the AI models in different contexts. Why? Please provide a tabular explanation. If not, the article does not have any interesting information for readers.

Thirdly, authors did not verify the biasness esteems to included articles, the selection process does not a clear information, while rigorous in terms of applying inclusion and exclusion criteria, may still be prone to selection bias. In methodology section, the work should have selection criteria for those 49 articles.

Lastly, The study uses a largely descriptive synthesis approach, summarizing the findings of individual studies without offering much critical insight into how these studies' results relate to each other. Why? Justify it. If not, discuss the relations among results of all those studies.

Author Response

We sincerely thank the reviewer for their thoughtful feedback and valuable suggestions. Your insights have helped us improve the clarity, quality, and overall impact of our work. 
Kindly see the attached file for detail response. 

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Accept in present form

Author Response

Thank you so much. We appreciate the time and effort you dedicated to reviewing our manuscript.