Robotics Perception: Intention Recognition to Determine the Handball Occurrence during a Football or Soccer Match
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis article investigates methods for handball recognition in football matches, which has certain novelty and practicality. However, there are still some issues as follows:
1. The abstract should provide a more comprehensive summary of the research content, and the conclusion section also needs further clarification.
2. The summary of related work is not comprehensive enough and should be supplemented.
3. Figure 8's text is too small to read clearly. Any other figures with small text in this article should be optimized. Additionally, explanatory text can be added to the figure titles.
4. The recognition speed should be specifically indicated.
5. Please list the camera distance, camera pixels, and field of view angle.
6. If a single camera is blocked (which is common in matches), will the method become ineffective? Furthermore, what methods can be adopted to address this? Please provide explanations.
7. In the experiments of this article, the recognition accuracy of the adopted technology is high in small indoor backgrounds. However, if applied in actual football field environments where backgrounds and recognition ranges are larger, is the method still effective? The authors should clarify this.
The method adopted in this article is relatively simple. The limitations of the technology used in this article should be explained, and some prospects should be provided to guide future research.
Comments on the Quality of English LanguageThe overall English language proficiency demonstrated in this paper is relatively high. The author employs a rich vocabulary and diverse sentence structures, lending the paper a certain linguistic elegance. Throughout the exposition, the author's expressions are clear and the logic is rigorous, facilitating readers' understanding of the paper's main ideas and viewpoints. However, there are occasional grammar errors and instances of imprecise wording, which may cause some confusion for readers. Additionally, in certain paragraphs, the language appears slightly verbose, and could benefit from simplifying sentence structures and condensing wording to enhance the flow and readability of the article.
Author Response
Please see the attachment to get the answer. Thank you.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this paper, authors presented the robotics perception and intention recognition in handball events during football or soccer matches. To further improve the quality of the paper, the following suggestions are put forward:
1. In this paper, Figure 2 provides too little information and is not prescriptive. The text in Figure 8 is too small to be clear.
2. What are the differences and advantages of the research method proposed by the author compared with the existing research
3. In the actual football match, the light is very complicated, and the background contains many interference images. How does the method proposed in this paper deal with these problems to ensure the recognition accuracy.
4. What is the recognition time of the recognition algorithm proposed in this paper? How does the paper consider the challenges of real-time identification and possible latency issues during live football or soccer matches.
Author Response
Please see the attachment to get the answer. Thank you.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsFrom the title, it is understood that the application goes as far as detecting the touch with intention of the ball with the hand. From the article, the results mainly relate to the detection of objects: hand and ball.
The title needs to be better linked to the content. Why “Robotics Perception”? What are the obtained results regarding “Intention Recognition”?
The interesting aspect of an application with a high degree of originality would even be the automatic signaling of the violation of the game rule. The achievement of objectives is discussed in a general way, without clear and concrete aspects of an implemented application.
The article refers to synthetically created test images. A clear answer is needed as to why this necessity arises since there are so many real images that can be used.
The article needs to be reorganized in order to make a clearer distinction between the general objectives set by the collective and those actually achieved and presented in this paper.
Another distinction must be made clearly between what is identified as a general strategy to detect violations of the rules (described in words) and what is technically implemented through equipment and programs (description in algorithms, technical schemes and programs).
It is necessary to make a clear distinction between the use of already existing programs and algorithms and the elements of original own contribution made by the authors of the article. The presentation and implementation of experimental data needs to be clearly improved in this case.
Comments on the Quality of English LanguageIn common parlance, football is a game where the ball is controlled with the foot, and handball is a sport where the ball is controlled with the hand. How to make the difference between the term “handball” related to the ball control with the hand during soccer play and the term “handball” related to the other sport?
Author Response
Please see the attachment to get the answer. Thank you.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe article presents a novel approach to improve the decision-making process in football matches by identifying handball events and their intentionality using computer vision and robotic perception. This study stands out for its application of object detection techniques and intention recognition to critically assess a player's actions during a game.
1. The authors provide a compelling rationale for the study, highlighting the importance of accurate decision making in football and the potential for technology to assist referees. However, the introduction could be improved by discussing the current limitations of existing technologies, such as VAR, in capturing the intention behind handball events.
2. The review of related work succinctly covers various approaches to analysing football games, with a particular focus on computer vision applications. To strengthen this section, the authors could consider including a discussion of previous attempts to identify intention in sports, providing a clearer distinction of how their approach contributes to the field.
3. The proposed methods are well-articulated, with a clear breakdown of the process of object detection, intention recognition, and event definition. The inclusion of equations and model descriptions adds depth to the understanding of the approach. However, a more detailed discussion on the choice of technologies and algorithms (Detectron2, Kalman filter, etc.) and their comparative advantages would be beneficial.
4. The results are promising, with high accuracy rates in object detection and intention recognition. Visual aids and tables enhance the clarity of the presented findings. However, the discussion of the limitations encountered during the experimentation, particularly around gaze detection and environmental factors affecting intention recognition, is crucial and needs to be expanded.
5. The discussion effectively highlights the potential impact of the study on improving fairness in football through technology. However, it would be enriched to include considerations on the practical implementation of this technology in live matches, including the necessary infrastructure and referee training.
6. The authors outline meaningful directions for future research, especially on data set expansion and multi-camera setups. Expanding on how intention recognition can be adapted to other ambiguous game situations could further underscore the broader applicability of the research.
The manuscript presents a compelling approach to a significant issue in football refereeing. However, further refinement is needed to address the literature review gaps, methodological justification, practical implementation considerations, and a more detailed exploration of the limitations encountered.
Comments on the Quality of English LanguageThe manuscript is well written, with minor editorial corrections needed for clarity and flow. A thorough proofreading by a native English speaker is recommended to polish the text further.
Author Response
Please see the attachment to get the answer. Thank you.
Author Response File: Author Response.pdf
Reviewer 5 Report
Comments and Suggestions for AuthorsThank you for your submission and your efforts in writing this article. Please consider the following comments and recommendations as constructive feedback with the only purpose of improving the quality of the current research.
The research topic introduced in this paper is interesting, valuable for the artificial intelligence (AI)’ community, and in line with the journal’s aims.
This version of the paper highlights the results of an empirical study using well-established tools to detect when a football player touches the ball intentionally.
The paper objective is interesting, but it contains too much major issues to be accepted in a journal. Moreover, the authors re-use entire sections from an already published article.
Reasons to Accept
- The objective is interesting.
Reasons to Reject
Major issues:
- In Section 3 Proposed Methods, the definition of the handball event is presented, specifically in Figure 1. In that figure, how the decision step “Is the hand makes body bigger?” (which contains grammatically errors) works? When that decision state is activated?
- In Section 3.2.1. the definition of intention as an action or an emotion need a reference, given that in different research areas (software agents, cognitive science, and many branches of social sciences) there is a clear differentce between what is an intention, an action, and emotions. Suggestions: clarify those definitions based on academic, well-established references; remove the sentence “In this work, we need a definition of what intention is.”; explain where the framework for decision-making presented in Figure 2 comes from; revise the English grammar (use a software for that) from lines 110-125; use academic writing in the entire Section 3; add a reference or specify the rationale why using the sight is possible to determine the intention of a person (lines 126 and 127).
- In section 3.2.2. and 3.2.3, the authors establish the variables to determine the action of the (remove double the) player. Then, it is presented a methodology in Figure 5, and equations to detect times and probabilities to calculate whether a player intentionally touches or not the ball. This section have not only grammar and typos problems, besides in line 149 says that the “…player’s speed was almost zero because the player was not moving.”, which is confusing.
- The article presents too many grammar issues, typos and non-academic language. Some parts of the article are very difficult to understand. Suggestion, do not submit an article without using a software for checking these issues.
- One of the aims of the article (line 58,59,60) is “foster a equitable surroundings in the gameplay”, how the identification of certain handball events can foster equitable surroundings in a football game?
- Remove duplicate paragraph from lines 165 to 168.
- Section 3.3.2. is almost a copy from the article reference [25] in the article. Moreover, the authors do not explain some of the selected settings for using such algorithm. Furthermore, from lines 186 to 193, there are several terminology that has not been introduced previously, which comes directly from [25].
- Section 4.1. presenting the results is a copy of the results presented in reference [25]. It contains the same exact results, with the same plots.
- Section 4.2.1. and 4.3. introduce methods not results. These sections should be presented in methods, not as results.
- The Discussion and Conclusions sections are to weak, and they share too much similarities with the reference [25]
Comments on the Quality of English Language
o Lines 14,16,19 among others, please add a space before adding the citation, for example, adding ~\cite[ref]
o Revise the English grammar of the sententes: “There are lots of events and actions going on in a football match. Of these, the handball event is one of the important events.”. It sounds a common language expression but not an academic one.
o In lines 20-22 the handball event is defined, but it was used previously from lines 16 to 19. In academic articles, usually, key notions are defined first, then used.
o In line 33, the authors mention “The article…” but it is not clear what article! Usually, in academia, if an author wants to talk about one article writes: “In [CITE], the authors…”. Similar problem is in line 38, “In a separate investigation, researchers…” but the authors do not mention immediately the article. Instead, add the citation after you mention it: “In a separate investigation [CITE], researchers…”.
o In line 52 and 53, CNN, LSTM, and MLP are mentioned but not defined previously. Suggestion, add the full name as in lines 48 for HAR. Never assume that the reader knows what is the meaning of a acronym. Alternatively, add a acronyms list.
o Line 58, “intention” in the context of the article is not a “human emotion”. If that is so, please add a reference.
o Grammar issues (among many!):
§ Line 31, remove “has” for “have”
§ Line 59, remove "…a equitable…” for "…an equitable…”
Author Response
Please see the attachment to find the answers. Thank you.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors have carefully modified the content of the paper and the analysis of the results as suggested previously. The research route of the paper is clearer and the research results are more credible.
Author Response
Please see the attachment. Thank you.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsIt might be useful if from this experience the authors of the paper would have understood that the purpose of responding to the comments received is not to justify and explain to the evaluator what was intended, but to improve the paper.
The actual result of improving the paper must be the added / modified paragraphs and not the explanations sent to the evaluator.
Author Response
Please see the attachment. Thank you.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsImprovements:
1. The authors added a discussion on the limitations of VAR technology, addressing previous feedback to provide context for the need for their system.
2. The manuscript now includes a broader discussion of intention recognition in sports, which was previously recommended to better position your work within existing research.
3. Additional details on the choice of technologies like Detectron2 and the Kalman filter were included, helping to clarify why these were suitable for the study objectives.
4. The presentation of the results with visual aids and tables has been improved, which enhances the clarity of the findings.
Remaining concerns:
1. The discussion on limitations, particularly around gaze detection and environmental factors, was noted to need to be expanded. The revised manuscript provides some discussion of this, but could still benefit from more detailed exploration.
2. Although there is some discussion of how the technology could be implemented in live matches, the manuscript could further benefit from a deeper exploration of the necessary infrastructure and referee training to fully integrate this technology.
3. The manuscript outlines directions for future research, including the expansion of the data set and the use of multicamera setups. It might also be beneficial to elaborate on how intention recognition can be adapted to other ambiguous game situations.
Comments on the Quality of English LanguageThe manuscript is noted to be well written with minor editorial corrections needed. It is recommended to have athorough proofreading by a native English speaker to enhance clarity and flow.
Author Response
Please see the attachment. Thank you.
Author Response File: Author Response.pdf
Reviewer 5 Report
Comments and Suggestions for Authors
Thank you for your submission and your efforts in writing this article. Please consider the following comments and recommendations as constructive feedback with the only purpose of improving the quality of the current research.
This version of the paper has been improved considering the initial submission. The authors of the article have partially considered some of the review arguments made, such as the improvement in the grammar and typos, as well as adding references to some specific statements.
However, there are still some minor issues that need to be corrected, such as typos (see below), as well as there is a two major issues that I recommend take action.
Major issues:
- According to the authors, this research is an extension of a previous published article, that is why the initial confusion for self-plagiarism. However, in this version the authors do not mention that fact in the introduction, neither in other section. Without a clear acknowledgment stating something like "... that this article extends results from [XX]... " or similar, the current work may look like self-plagiarism. Then, please, add that sentence in the introduction, and in the Discussion section, highlight the additional experiments that were included in this version.
- In Figure 1, the flowchart that leads the methodology is seems to be flawed. The same comment was highlighted in the initial review. Please revise that graph.
- In some subsections of Results, the authors describe the methods and tools used. Those sentences need to be in the Methods section. This comment was highlighted in the initial review.
Minor issues:
- Duplicate "dataset" in line 68
- CNN model instead of "CNN-model" line 70
- "sports applications' perception" instead "sports applications perception" line 91
- There are more small typos that need to be fixed, please use an automated software to re-check the document.
Comments on the Quality of English Language
Minor issues:
- Duplicate "dataset" in line 68
- CNN model instead of "CNN-model" line 70
- "sports applications' perception" instead "sports applications perception" line 91
- There are more small typos that need to be fixed, please use an automated software to re-check the document.
Author Response
Please see the attachment. Thank you.
Author Response File: Author Response.pdf