All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
You carefully considered all reviewers' comments and the overall quality of the manuscript has, in my view, improved a lot.
Please, carefully follow the Reviewer’s remaining indications
I think the authors have generally done a good job content-wise, and am generally happy with the content of the text. There are still some language problems that needs to be fixed, but this is hopefully a minor revision.
The authors have generally answered to my concerns. A remaining comment: on row 123 it is stated that other keywords were used. Please list which they were.
The authors have generally answered to my concerns.
As I stated initially, I think the authors have done a good job on addressing my concerns content-wise. What remains is to go over the language again. I am not good at proof reading, so I will not provide a complete list, but I will point to places where I have noticed problems:
On row 43 "Where the movement ..." is an incomplete sentence. My guess is that this is actually meant to be part of the previous sentence. If not, it is still unclear which type of hand gestures you are considering (my understanding is in-air hand gestures, not on-device gestures like gestures on a touch screen).
The sentence on row 44 "Which gives an example...." is unclear - what does the "which" point to. I would suggest re-phrasing this sentence (as well as the one before) to make the intent clearer.
Row 63 "huge expand" - is this correct english?
Row 100 "intensity" of a study - what is this? I don't quite understand what you mean - please re-phrase
Row 126 the sentence about the duplicates is long - maybe split it to make it easier to read?
Row 136 is "unwell-organised" really good english? Maybe re-phrase (eg. poorly organised).
The text in 1.1 and 1.2 is quite technical, and not being an expert I find it harder to comment on.
In section 2 there is some phrasing which is dated - in the old days one would say deaf & dumb, because deaf persons might not speak (not seeing the point since they were unable to hear). I would suggest using persons with hearing or speech impairments (if this is what you mean). I also note that you say "sign language", but it appears what is actually meant is finger spelling? In a full sign language facial expressions play a part and you might use your hands on another part of your body - eg. touch your ear as to indicate earrings /show the shape of a breast on the chest to indicate "woman" (I am no expert on sign language, these are examples I have been told).
Section 3 again, is technical. The final two sections work for me, although the final sentence on row 762 could possibly be made less passive to make it flow better (something along the lines of "a common problem found in many of the studies was overfitting in the datasets").
Since this list is not complete, I would encourage the authors to make a general final round of language "polish".
Please, carefully consider all the issues raised by the reviewers and revise your manuscript accordingly.
[# PeerJ Staff Note: Please ensure that all review comments are addressed in an appropriate rebuttal letter, and please ensure that any edits or clarifications mentioned in the rebuttal letter are also inserted into the revised manuscript (where appropriate). Direction on how to prepare a rebuttal letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
This is a review paper, which aims at providing an overview of current tools and techniques within the field of hand gesture recognition. The paper is generally well referenced, but the text is somewhat hard to read, and I would recommend additional editing with a focus on making the text more readable. The authors provide a nice overview figure of their method, but it might be helpful if some figures could be added that also visualise the outcomes of the work. My main concerns, content-wise, with this work lies in the initial sections - the introduction and the methodology (see experimental design).
Introduction
I find it unclear what the authors really mean with hand gesture. Using gestures on screen has been a standard technology for a long time, an in essence already the mouse is an indirect hand gesture interface. The intro, on the other hand, is written as if hand gesture recognition is something that lies in the future.
I also question the general statement that users need to be experts; yes, gesture UI:s are often designed that way, but this could be seen as a design feature – if the system would be truly intelligent (at some point in the future) it might respond appropriately to user gestures the way humans are able to. Thus, this point needs some explanation and discussion.
The intro is written with image recognition as the main technology for gesture recognition. I would suggest it more appropriate to use more generic wording – after all the literature study does not focus on image recognition alone.
I must admit that I question that focusing on 2016-2018 will give us “the valid base of the current situation and technologies of hand gesture recognition.”. With this focus, standard technologies and well established tools will be missed. You also miss early innovative work that may have been discontinued due to technology immaturity or just trends in the technology field. As an example: in the related field of haptics, a lot of work was done in the 1990’s- early 2000’s, and then, since much of the technology development turned mobile, less work was done, leading persons who revisit the field to miss quite a lot of the existing knowledge if they only take the most recent work.
To conclude, I would like to see the introduction rewritten to:
1) Better motivate the very limited selection
2) Clarify what is meant by gesture
3) Improve the explanation of gesture UI:s
A note: the authors may want to check out the ISO standard 9241-960 Framework and guidance for gesture interactions.
Survey methodology
As already explained under introduction, I question the assumption that limiting the search to 2016-2018 will provide answers to the questions that are listed as the research questions. Either provide a better motivation for why this is the case, or change the questions so that the study answers the questions you have.
I am also slightly skeptical to the search terms. It is always a challenge finding the terms people use, and one may easily miss chunks of research simply because slightly different phrases/terms/expressions are used in an area. I would have liked to see an early exploration of which terms that could be used for the search. Eg. unless some exploration is done, you won’t know if some authors might talk “gesture interaction”, “gesture interfaces” leaving out the “hand” even though the actual interaction/system they discuss is hand based.
I would like more details on the process. It is stated “excluded for having titles with no relevance to the review in general” – please explain a little more what this actually means. The same goes for “All 244 papers were reviewed and evaluated in reference to the research questions, this process excluded 100 papers” – what was really excluded at this stage? Please explain better, maybe provide examples? I find it unclear what was really meant by “49 unwell-organized documents were removed due to lack of validation and weakness of results justification” – please provide more clarity on this process.
To conclude, this section needs revision and clarification. Please:
1) Improve the motivation for the research questions alternatively rewrite the questions themselves.
2) Provide more detail on the search and selection process. As it is, the reader cannot really follow what you did, and is thus cannot judge how appropriate your results are.
Results and analysis
Not being an expert in gesture recognition (I have used it, and have implemented basic gesture recognition but would not consider myself an expert), there is much in this section that appears quite useful. At the same time, the section is hard to read, and I wonder if it wouldn’t be possible to organize the text more also in the subsections (maybe provide a short overview first in each subsection, and possibly some illustrative images)? As I stated initially, I don’t have problems with the things that are found – these appear sound – what I am worrying about is what it is that might be missing, given my stated methodological concerns.
To conclude
1) Please re-work this section so that it becomes more readable.
To sum up, while this texts does provide a useful overview of tools and techniques in gesture recognition, my main concerns with this work is methodological. I question the limitation to 2016-2018, and I am not convinced the method actually answers the stated research questions.
The paper follows the appropriate structure for a review paper.
In the main, the English language is easy to read, however, it would benefit from proofreading particularly for grammatical errors. For example, (1) there are multiple examples of run-on sentences (e.g. line 79 “In artificial intelligence, machine learning enables the computers to learn without being explicitly programmed (Margaret, 2016), there are two types of learning; the process when algorithms ….), (2) incomplete sentences (e.g. line 82: “Classification aims of building a model based on previous training hand gestures to classify new hand gestures.” and (3) poor/incorrect wording, including use of the word “researches” where “research” would be more appropriate (e.g. in the abstract’s Conclusion “We shall also introduce the most recent researches..” and reference to “unwell-organized documents” in line 122 and elsewhere.
The raw data is supplied as a link to a .rar file, which I did not seem to be able to open without downloading and installing additional software.
This paper presents a review of prior work on hand-gesture recognition. The topic is timely and interesting, however I feel the survey methodology has a number of important limitations:
1. The search was limited to IEEE Explore, and hence would have missed a large body of related literature outside of IEEE, for example, papers in the ACM Digital Library.
2. Two main “keywords” were used: “hand gesture recognition” and “hand gesture techniques.” It’s unclear if these search terms would have found papers that use other related terminology (e.g. in-air gestures) or terms that describe some of the common technologies (e.g. Leap motion). It is also important to clarify if touchscreen gestures are included in this review or not.
3. The methodology needs further elaboration and examples of exclusion criteria, for the reader to be able to understand fully and have confidence in the method. For example “after removing duplicates” – how were duplicates defined? “papers were excluded for having titles with no relevance to the review” – what were the criteria for relevance, who applied them, are they unambiguous and repeatable by different reviewers? This also calls into question the appropriateness of the search terms (see #2 above). “unwell-organized documents were removed” – again, how was this defined and applied?
It would also be good to see some further explanation of how many researchers reviewed the final list of included papers and what were the steps involved in reviewing the papers.
I also wonder if “scoping review” or “state of the art review” would be a more accurate reflection of this work, since “systematic review” often suggests a review of studies that have a common aim of answering a (healthcare-related) question.
The authors present a substantial body of literature. To a large degree, it reads as a list of who did what and it would be more valuable to the reader if there was greater synthesis of the literature, drawing out insights derived from having oversight of the field.
The authors have grouped the results/analysis into headings and subheadings, which is important for a review. However, I felt the chosen structure of headings and subheadings did not work very well, particularly for section 1 and its subsections. For example, the choice of features depends very much on the underlying data that has been captured – and so it does not seem to make much sense to list features used by computer-vision based systems alongside EMG features, without the context of what each system is aiming to achieve. One suggestion is to organise the literature in terms of the acquisition systems, and then present the features and classification methods that have been investigated for each acquisition system. This would enable like-for-like comparisons to be made, and also enable comparisons across different acquisition systems. Similarly, the technical challenges are highly-dependent on the type of system being used, and so again, it would make more sense to present the challenges by system, and then draw out common themes within and across types of system.
The section on applications presents a very limited view of where gestures are being used and explored currently, and I suspect is a result of the search being limited to IEEE Xplore. It does not seem to reflect the current state of the art, and I think it likely that expanding the search would enable the authors to expand out the “Others” section quite substantially.
The section on “the future of hand gesture recognition” is quite speculative, and while I understand speculation is acceptable and welcomed in PeerJ, I feel it still needs a good grounding in the literature and this grounding is not evident in this section.
The conclusions don’t all seem to follow from the literature – for example, much of the literature presented in the paper appears to use camera sensors, whereas the paper concludes that surface EMG “were the most acquisition tools used”.
The topic of this paper is timely and would be of interest and value to the research community.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.