Jump to content

Talk:WikiJournal of Science/Multiple object tracking

Page contents not supported in other languages.
Add topic
From Wikiversity

WikiJournal of Science
Open access • Publication charge free • Public peer review • Wikipedia-integrated

WikiJournal of Science is an open-access, free-to-publish, Wikipedia-integrated academic journal for science, mathematics, engineering and technology topics. WJS WikiJSci Wiki.J.Sci. WikiJSci WikiSci WikiScience Wikiscience Wikijournal of Science Wikiversity Journal of Science WikiJournal Science Wikipedia Science Wikipedia science journal STEM Science Mathematics Engineering Technology Free to publish Open access Open-access Non-profit online journal Public peer review

<meta name='citation_doi' value='10.15347/WJS/2023.003'>

Article information

Submitting author: Alex O. Holcombe[a][i] 
Additional contributors: Wikipedia community

See author information ▼
  1. School of Psychology, The University of Sydney
  1. alex.holcombe@sydney.edu.au

 

Plagiarism check

Pass. Report from WMF copyvios tool flagged some false positives (not regarded as plagiarism) due to common stock phrases like "number of objects tracked" that repeated multiple times. OhanaUnitedTalk page 22:35, 10 November 2022 (UTC)Reply

First peer review


Review by Liqiang Huang , The Chinese University of Hong Kong
These assessment comments were submitted on , and refer to this previous version of the article

This is a very good summary of the studies of Multiple object tracking. The author is an active researcher on this topic. This review article is clear, precise, adequate, and up-to-date. I only have two suggestions.

1. The work of Yantis (1992, Cognitive Psychology) should be discussed. This study demonstrated that the tracking performance critically depends on whether the targets can be maintained as a coherent but nonrigid virtual polygon. The tracking is significantly disrupted when that visual interpretation is violated (e.g., a vertex of the virtual polygon crosses over an opposite edge of the polygon).

Response

Thanks to the reviewer for pushing me to revisit this paper. To accommodate it and related work, I have added a new section on grouping.

2. The work of Franconeri, Jonathan, and Scimeca (2010; Psychological Science) should be discussed. This study argued that MOT is limited only by object spacing (how close together objects can be), but not by speed, time, or capacity. This opinion is somewhat extreme and may not be completely right, but it has made an important point about the role of object spacing.

Response

Object spacing's effects speaks to spatial interference, which is addressed in the "Spatiotemporal limits" section of the manuscript. The Fraconeri et al study, but also many related follow-up studies, have contributed to our understanding of spatial interference, with extensive discussion provided in the book of mine cited in that section. Combining the findings of the follow-up studies, as the reviewer may have been hinting, indicates that the strong Franconeri, Jonathan, & Scimeca (2010) claim is wrong, and the present manuscript cites the work that overturned that conclusion and supported the new, more nuanced understanding. As a result of that later work, it no longer seems important to me to discuss that paper explicitly in a Wikipedia article, as Wikipedia articles prioritize the historical development of theories even less than a scientific article might (unless the Wikipedia article is actually about that theory). I certainly could discuss this specific paper, and in fact I have explicit discussion of it in my book that I cited, but I think such discussion would be only of historical interest (as an ambitious hypothesis that was subsequently disconfirmed) so it is my opinion that it is not a judicious use of space in this proposed Wikipedia article.

Second Peer Review

reviewer-annotated pdf file.
reviewer-annotated pdf

Review by Todd S. Horowitz , National Cancer Institute
These assessment comments were submitted on , and refer to this previous version of the article

This is a good, accurate and balanced summary of the MOT literature. I think it does a fine job of indicating what we know, and what the major issues in the field are, without pushing the author’s own perspective. I do have some suggestions for improvement.

(1) I think that the second paragraph in the procedure section should be expanded to describe major variations on the MOT method. For example, experiments where stimuli do not move randomly but rotate along a common circular path are quite common; without a description of this type of study, it is difficult to understand the author’s explanation of temporal limits in tracking.

Response

Thank you for this important point. I have rewritten the end of the Procedure section to address this, adding a few sentences.

(2) The description of the “shell game” study (Pailian et al.) refers to the participants as Harvard undergraduates, but information about the study population is not provided for most of the studies described here (except in the “human variation” section), so it’s a little jarring. I suggest deleting this description, and instead adding into the “human variation” section a brief discussion about the fact that almost everything we know about MOT has been derived from WEIRD populations (typically undergraduates, though not all at Harvard).

Response

As suggested, to address the jarring mention of "Harvard undergraduates", I deleted the phrase. To address the issue that the literature uses almost exclusively WEIRD participants, I added some sentences in the middle of the "Human variation" section, which I've now re-named "Human variation and development". To increase the representation of groups besides Western university undergraduates, I reported results from studies of children and of persons with autism spectrum disorders, using papers that include those mentioned by Reviewer 2 in his comment 3.

(3) While we’re on the topic of individual differences, I think it would be worth including work on clinical populations, such as Kirsten O’Hearn’s work on Williams syndrome (O’Hearn, K., Hoffman, J. E., & Landau, B. (2010). Developmental profiles for multiple object tracking and spatial memory: Typically developing preschoolers and people with Williams syndrome. Developmental Science, 13(3), 430–440. https://doi.org/10.1111/j.1467-7687.2009.00893.x

O’Hearn, K., Landau, B., & Hoffman, J. E. (2005). Multiple object tracking in people with Williams syndrome and in normally developing children. Psychological Science : A Journal of the American Psychological Society / APS, 16(11), 905–912. https://doi.org/10.1111/j.1467-9280.2005.01635.x) and autism (Koldewyn, K., Weigelt, S., Kanwisher, N., & Jiang, Y. (2013). Multiple object tracking in autism spectrum disorders. Journal of Autism and Developmental Disorders, 43, 1394–1405). Also, to the discussion of aging effects, add developmental studies (e.g., Trick, L. M., Jaspers-Fayer, F., & Sethi, N. (2005). Multiple-object tracking in children: The “Catch the Spies” task. Cognitive Development. http://linkinghub.elsevier.com/retrieve/pii/S0885201405000249; Cheng, C., Kaldy, Z., & Blaser, E. (2019). Two-year-olds succeed at MIT: Multiple identity tracking in 20- and 25-month-old infants. Journal of Experimental Child Psychology, 187, 104649. https://doi.org/10.1016/j.jecp.2019.06.002).

Response

I really appreciate the reviewer's mention of two papers on Williams Syndrome, which I did not know about and have now added.

(4) In the section on “Interference with concurrent performance of other tasks”, I think it miught be worth mentioning my paper with Ian Thornton (Thornton, I., & Horowitz, T. (2015). Does action disrupt Multiple Object Tracking (MOT)? Psihologija, 48(3), 289–301. https://doi.org/10.2298/PSI1503289T) showing that action does not disrupt MOT, especially given the later mention of multiple object avoidance, which also involves concurrent action.

Response

Reading the Thornton & Horowitz (2015) paper mentioned was helpful, but I wasn't sure how to describe the results in a way that fit well with my manuscript. The reviewer's comment suggested to me that there was no cost to action for MOT – although the authors did find that the dual-task cost was statistically significant, the authors characterized it as small. They're probably right about that, but it would require some subtle reasoning to justify a simple conclusion such as that action does not disrupt MOT, and I'm not sure how to make that argument for a Wikipedia audience. Note that for my book, which some of the points in the present manuscript is based on, I did start to draft a chapter on dual-task interference but I abandoned it when I realized the book would be too long with that included, and as a result I don't feel I have enough mastery of the dual-task literature to go deeply into it. Also, the recent paper by Terry and Trick seems to paint a more complicated picture (although I haven't read the paper in detail).

(5) I think it would be worth spending a little more time on the theories and models. You mention the original FINST model in the intro, and the Lovett et al. IMOT model. Under “theories and models” What about the FLEX model (Alvarez, G. A., & Franconeri, S. L. (2007). How many objects can you track? Evidence for a resource-limited attentive tracking mechanism. Journal of Vision, 7(13), 14.1-10. https://doi.org/10.1167/7.13.14)? Vul et al.’s particle filter model (Vul, E., Frank, M., Tenenbaum, J., & Alvarez, G. A. (2009). Explaining human multiple object tracking as resource-constrained approximate inference in a dynamic probabilistic model. Advances in Neural Information Processing Systems, 22. http://books.nips.cc/papers/files/nips22/NIPS2009_0980.pdf)? I don’t think you need to provide a complete guide to all the models, but just to list one seems odd

Response

I appreciate the reviewer prompting me to add more about models - this is a complex topic and I'd shied away from it because I haven't seen a lot of critical tests comparing models, instead people tend to use pretty flexible models that can explain the same things, for example slots-plus-averaging can explain gradual performance decreases with display parameters, but this is rarely grappled with in the literature – the MOT literature is far behind the visual working memory literature on this. I've added a mention of additional models and a description of some of the issues they get at.

(6) While I’m very proud of the “Tracking Unique Objects” paper, Multiple Identity Tracking (both the paradigm and the term) was originally introduced by Oksama & Hyönä (Oksama, L., & Hyönä, J. (2004). Is multiple object tracking carried out automatically by an early vision mechanism independent of higher-order cognition? An individual difference approach. Visual Cognition, 11(5), 631–671.).

Response

I left out the Oksama & Hyona 2004 paper for brevity, although I do cite that paper elsewhere in the manuscript.

See also comments on the pdf.

Response

Reviewer 2 PDF annotation 1. The reviewer describes the explanation of the dissociation between tracking and motion direction judgments as being due to dedicated motion system mediating direction judgments as "perilously close to redescribing the data". I think there are other ways besides a dedicated motion system that it could be explained, and which some might favor if we didn't already have as much evidence for a dedicated motion system as we do from several decades of motion perception research. For example, early researchers (and naive Wikipedia readers) might have expected that motion would be detected based on change in position, using an actual position tracker that is not separate from position perception, rather than using dedicated detectors such as Reichhardt suggested. In that case, the dissociation would be much more surprising, and for naive readers of Wikipedia who don't know about the motion aftereffect and other independent evidence for detectors of motion, they might hesitate to conclude there is a motion system distinct from that used for tracking extended trajectories, without the guiding hand of the sentence I included. Still I appreciate the reviewer's comment as it pushed me to add more substance; I have revised the sentence to the following: "This dissociation between motion perception and object tracking is thought to reflect that direction judgments can be based on low-level and local motion detectors responses that don't register the positions of objects."

Reviewer 2 PDF annotation 2. About the circular array display my manuscript refers to, the reviewer writes "Needs to be clearer that we've moved from the classic Pylyshyn style independent, random object motion to a rotational paradigm. Maybe briefly list the kinds of MOT paradigms above?" I want to avoid having to try to come up with a taxonomy of the wide variety of displays people use, which is somewhat implied by a list, instead I'm hoping to maintain writing the article in a way that implies an infinite range of possible trajectories could be used. So to better convey that, what I've done is add to the second paragraph under the Procedure subsection. I also added a new sentence to introduce the circular display before the sentence commented on by the reviewer.

Reviewer 2 PDF annotation 3. Thank you for pointing out that the meaning of that sentence is obscure. I'm not sure I should get into specifics of the experiment as my manuscript has gotten pretty long, but I hope the following is clearer – I think it is clear enough that there is a dissociation and the reader should consult the cited paper to understand the nature of that dissociation: "While one might expect this to tap into the same processing as an MIT task, the relationship between the two is unclear, as there is evidence that attentional tracking occurs can occur along a different trajectory than that which is the basis of updating the memory of an object's features."

Reviewer 2 PDF annotation 4. The reviewer pointed out a very unclear part of a sentence, "the corresponding dimension", so I've simply deleted that phrase. It did mean something but it's unimportant for this article; it referred to a detail that a motivated reader can learn about by reading the cited paper.

Treenxorr (discusscontribs) 06:48, 23 January 2023 (UTC)Reply

Response by author

Dear Dr. Leung,

Thank you for coordinating the process and thanks very much to the reviewers for their comments. Before detailing how I have revised the manuscript to address the reviewers, I would like to raise a question that you may be able to help me with. Because in the field of engineering and computer vision, the phrase "object tracking" refers to tracking by computers, I think the Wikipedia page should start with something like "For visual tracking of objects by computers, see Video tracking." I would appreciate your advice on whether you think that is appropriate and if so, exactly how I should format such a phrase.

The edits I made in response to the reviewer's comments have substantially lengthened the manuscript. This made it important to add additional structure, so I've found some ways to make the article a bit more hierarchical rather being restricted to top-level headings. To do this, I demoted the heading "Use in ability testing and training" to make it a sub-heading of "Human variation and development". By bulking up the Theories and Models content, I was also able to divide that section into "Serial or parallel?" and "Slots or resources?" subsections.

I have also made a lot of small edits to improve the writing of most of the sections.

Thanks again to the reviewers. Their comments have sparked many improvements to this candidate Wikipedia article.

Aoholcombe (discusscontribs) 23:57, 23 February 2023 (UTC)Reply

Hi Alex. Thank you for the comments and revisions in a timely manner. We will leave the "for visual tracking of objects by computers..." link out of your manuscript as your submission is intended to be a standalone article that can be dual-published in Wikipedia as well. If the article gets accepted, we will generate a PDF based on your manuscript version. When we import your manuscript into Wikipedia, that link will be added back to the top of the Wikipedia page. We will now provide your revised manuscript to the reviewers for second readthrough. OhanaUnitedTalk page 18:16, 28 March 2023 (UTC)Reply

Editorial comments


Comments by Andrew Leung ,
These editorial comments were submitted on , and refer to this previous version of the article

The editorial board has a few comments and suggestions:

In the lead section:

  • "a subset are designated as targets." => "a subset are designated as targets and the rest as distractors."
  • "awareness of features such as color and shape are disrupted" => "awareness of features such as color and shape is disrupted"

In Procedure:

  • " the targets are indicated by being presenting them initially" => "the targets are indicated by being presented initially"
  • "is an indicator whether" => "is an indicator of whether"

In Spatiotemporal limits:

  • "the objects will be difficult to identify" => "the objects are difficult to identify"

In Updating of features other than position:

  • "Zenon Pylyshyn" => "Pylyshyn" as he has already been introduced.
  • "The term "object files" refers to the idea that" => "The concept of "object files" is that"

In Difficulty tracking unusual objects and object parts:

  • "vanMarle and Scholl" => "Kristy vanMarle and Brian Scholl"

In Neural basis:

  • "activation of areas of the parietal cortex increase" => "activation of areas of the parietal cortex increases"

In Human variation and development:

  • "The scores of a person will usually be similar" => "The scores of a person are usually similar"

In Slots versus resources:

  • "FINST theory first described in 1988 " => "FINST theory" as this has already been introduced.

OhanaUnitedTalk page 04:17, 4 April 2023 (UTC)Reply

Response

Thank you to the editorial board for pointing out all of those writing issues! I have gone through and fixed each of them. In doing so, I noticed that some of the associated sentences were generally awkward, so I re-phrased those to further improve things.

--Aoholcombe (discusscontribs) 20:43, 6 April 2023 (UTC)Reply