Evolution and Trends in Sign Language Avatar Systems: Unveiling a 40-Year Journey via Systematic Review

Comment

Action

In the manuscript entitled “Evolution and Trends in Sign Language Avatar Systems: Unveiling a 40-Year Journey via Systematic Review” the authors did a thorough and extended review of sign language avatar. The work is interesting and beneficial to research community. I would like to congratulate the authors for their effort.  The authors need to consider the following minor issues.

Thank you for your positive remarks on our manuscript. We appreciate your recognition of the thoroughness and potential value of our work to the research community. We are eager to address the minor issues you've pointed out and will take each of your comments into careful consideration to further enhance the quality of our paper.

1.    As the affiliation of both authors is same some they can use “1” as superscript with their names.

Fixed.

2.     Figures are readable but it is recommended to increase the text size in figures.

Figure 1 fixed.

Figure 2 fixed.

Comment

Action

This paper presents a comprehensive review of the existing literature on avatar systems, focusing on the technologies used in the creation of Sign Language (SL) avatars, the methodologies and metrics used in their evaluation, and the bibliometrics related to these avatars. The literature reviewed in this study dates back to 1982. The results of this comprehensive review include an analysis of the sign synthesis techniques included in the selected works. It also examines the corpora used in the development of SL avatars, as well as the design and animation strategies and methodologies adopted. In this context, this review illuminates current research trends in the field of sign language avatars, with the expectation that this comprehensive research will make a meaningful contribution to existing scholarship. I kindly recommend acceptance of the manuscript.

Thank you for your positive remarks on our manuscript. We appreciate your recognition of the thoroughness and potential value of our work to the research community.

Comment

Action

The authors present a systematic review of the literature on the evolution of avatar-mediated sign language. The reviewer acknowledges the authors' effort in analysing a vast literature and reporting so many metrics to best describe it.

The reviewer suggests minor revisions to be made to the paper before proceeding with publication:

Thank you for your positive remarks on our manuscript. We appreciate your recognition of the thoroughness and potential value of our work to the research community. We are eager to address the minor issues you've pointed out and will take each of your comments into careful consideration to further enhance the quality of our paper.

- in general, the paper turns out to be very very long. I suggest summarising any sections that do not report significantly new or relevant information. The abstract should be summarised to best highlight the work done and the main outcomes of the literature review conducted.

We understand the importance of a concise abstract that effectively highlights the key aspects of our work. We have revised the abstract to succinctly emphasize the primary objectives, methodology, and outcomes of our systematic review. The revised abstract is as follows:

Sign language (SL) avatar systems aid communication between the hearing and deaf communities. Despite technological progress, there's a lack of a standardized avatar development framework. This paper offers a systematic review of SL avatar systems spanning 1982 to 2022. Using PRISMA guidelines, we shortlisted 47 papers from an initial 1765, focusing on sign synthesis techniques, corpora, design strategies, and facial expression methods. We also discuss both objective and subjective evaluation methodologies. Our findings highlight key trends and suggest new research avenues for improving SL avatars.

- Table 1 is not very intuitive. I suggest that the authors state in full in the text the rationale behind the scientific questions that this review aims to answer.

Thank you for pointing out the issues with Table 1. In response to your comment, we have revised and converted the table to a more intuitive format.

The explanation of the background is also very long. I would suggest summarising.

Sub-sections 2.1 and 2.2 were summarized.

- In section 4.2.1, what does the analysis reveal then? The authors should better argue what the implications of their results are. For example, the authors say they identified 7 clusters of keywords. What does this imply? Add a comment.

Thank you for your valuable feedback on Section 4.2.1. We acknowledge the need to provide more context and implications based on our keyword analysis results. Here's our revised answer and the updated section:

Updated Section 4.2.1              SRQ1a: Keyword analysis

Keyword analysis was performed using VOSviewer 1.6.19 [86], a text-mining technique that constructs co-occurrence networks of keywords. A threshold of 2 was instituted, implying keywords must appear a minimum of twice in the bibliometric analysis for inclusion. After a meticulous manual data collection, 27 keywords out of an initial 150 were shortlisted for the final analysis. The co-occurrence network derived from these keywords, as depicted in Figure 3, unfolds into 7 distinct clusters, showcasing varying degrees of keyword interrelations. Within these clusters, 'sign language' prominently stands out as the pivotal keyword, accentuated by the size of its circle. This revelation indicates the diverse thematic concentrations present in the realm of sign language avatars, emphasizing the multifaceted and interlinked nature of research in this domain. For a more granular understanding, Table 8 delineates the frequencies of top keywords and their proportional representation among the chosen studies.

- In section 4.2.2, to talk about publication growth, I would suggest the authors calculate the trend of publications over the years. This would be a quantitative value of the actual increased scientific production.

In response to your comment, we computed the compound annual growth rate (CAGR) for the publications from 1982 to 2022. The calculated CAGR provides a more quantitative representation of the overall growth trend in the research domain. This value, combined with the graphical distribution of the studies over the years, provides a more holistic understanding of the research trajectory.

Updated Section 4.2.2 SRQ1b: Trend of published papers per year:

Figure 4 displays the distribution of studies spanning the years 1982 to 2022. From this distribution, an ascendant trajectory in the aggregate number of papers released in both conferences and journals becomes palpable. A notable peak is observed in 2022, marking the zenith of published papers within this timeframe. Delving deeper into the annual publication metrics, we notice variability: conferences consistently outpace journals in terms of publication frequency. To illustrate, 2013 saw four papers, while three papers were disseminated during 2007-2009, 2011, 2016-2017, and 2021. In parallel, two papers were promulgated during 2002, 2006, 2014-2015, and 2020. On the more modest end of the spectrum, solitary publications were recorded in 1999, 2004, 2005, 2010, and 2012. It's worth noting that select years, namely 2001, 2003, 2018, and 2019, registered a complete absence of pertinent publications.

Considering the above distribution and the calculated compound annual growth rate (CAGR) of 3.29%, the domain's evolution is marked by palpable expansion, culminating in a pronounced upswing in 2022. The fluctuations in annual publication numbers potentially reflect oscillating research engagements, dictated by shifting trends, technological advancements, and recalibrated research imperatives.

- I suggest the authors transform Tables 9-10-12 into pie charts to make them easier to read.

Table 9 converted into pie chart Figure 5

Table 10 converted into pie chart Figure 7

Table 12 converted into pie chart Figure 9

- The authors should comment in more detail on the results obtained. For example on the study in section 4.2.4, what can we deduce from the occurrence of papers per country?

We have revised section 4.2.4 to provide a more detailed commentary on the results and their implications.

4.2.4     SRQ1d: Country wise analysis 

Figure 5 offers insights into the geographical distribution of papers concerning SL avatars. Evidently, Spain stands as a significant hub for this research, accounting for 15% of the analyzed publications. The United Kingdom, Greece, the United States, and India also present considerable contributions, each constituting 9%. This dominance by Spain, coupled with notable contributions from other European nations, underscores the concentration of SL avatar research within the Western region. This could be at-tributed to the presence of academic institutions or research clusters focusing on sign language technology in these regions. It is also noteworthy that despite the vastness and diversity of Asia, represented countries from this region collectively contribute to only a fraction of the global research output, i.e., 11 out of 47 papers. Similarly, the Middle East and Africa's representation is notably sparse, with only three countries between them. Such observations indicate a potential gap in research across different parts of the world. It could signify limited resources, fewer research initiatives, or perhaps different research priorities in these regions. It might also suggest that collab-orations, funding, and support for SL avatar research are richer in Western and Euro-pean regions, thereby leading to a higher volume of publications. Figure 6 further de-tails the collaborative efforts among countries in this domain. The clusters formed, primarily by Western countries, highlight an active exchange of knowledge and col-laborative research endeavors within these regions. The collaboration between the UK, the Netherlands, Switzerland, and the USA can be a testament to the shared research interests and possibly institutional partnerships.

As a result, the country-wise analysis not only maps the distribution of research output but also points towards potential areas or regions where further research im-petus is needed. It also underscores the importance of fostering international collabo-rations to ensure a more holistic and diverse advancement in the domain of SL avatars.

- Figure 5 is very large for only 4 points.Can the authors use a more compact representation?

Fixed.

- What is the purpose of the study in 4.2.5? The authors should better justify this.

The main objective behind the author-wise analysis is twofold: first, to acknowledge and identify leading scholars and their substantial contributions in the domain of SL avatars; and second, to map out the intricate web of collaboration among these researchers, providing insights into collaborative trends and potential academic networks.

4.2.5 SRQ1e: Author-wise Analysis

Understanding the major contributors in any academic domain can provide key insights into its foundational works, pivotal research areas, and future directions. An author-wise analysis serves as a tool to pinpoint these significant scholars and comprehend the relationships and collaborations that might be shaping the field's discourse. Figure 7 showcases the standout authors, determined by the volume and impact of their contributions to the existing body of literature on SL avatars. Cordoba R., Ferreiros J., and San-Segunda R. have marked their dominance with each producing 4 critical works. Other noteworthy figures include Montero J. M., Pardo J. M., Efthimiou E., Fontinea S.-E., and Ebling S., with each attributed 3 important publications. To further elucidate the landscape, Figure 8 depicts the intricate collaboration networks among these leading authors. The visualization presents two prominent clusters. The first cluster encompasses Cordoba R., Ferreiros J., Macias-Guarase J., Pardo J. M., and San-Segundo R., signifying their close academic affiliations. In contrast, the second cluster associates Barra R., D’Haro L. F., and Montero J.M. These patterns not only highlight mutual research interests and methodologies but also hint at potential centers of academic excellence in the domain of SL avatars.

- There is a lack of an overall discussion section of the results, which the authors have included in the Conclusions. furthermore, the considerations expressed in line 773 "Future research ... " to the end of the paper, are very interesting and would require a special section according to this Reviewer that could be called "Take home messages and future direction".

we have separated the discussion of our results from the conclusion and included a distinct section titled "Discussion". Moreover, based on your suggestion, we have now added a separate section named "Future Direction" to ensure that the implications and prospective lines of inquiry are presented coherently.