Area Chair Guidelines

Contents

• Introduction
• Desk rejections of papers
• Evaluation of received reviews
• Discussion period
• Making a decision

Introduction

These guidelines are inspired by CVPR’2024 Author Guidelines. BMVC Programme Chairs adapted these guidelines to achieve the goals of BMVC 2024.

Thank you for volunteering your time to serve as area chair for BMVC 2024! As an area chair, you will play a crucial role in helping us maintain the high reviewing standards of the conference. We expect you to have at least two years’ experience as a postdoc or faculty, and to regularly publish in the lead computer vision venues (e.g. Google Scholar’s top 20 venues for computer vision and pattern recognition). You will work alongside the Program Chairs to make BMVC 2024 a success and help shape the future of computer vision research.

Please note we expect all ACs to take a proactive role in ensuring papers receive high-quality reviews and meta-reviews. Therefore, you must commit to being available and responsive during the relevant parts of the review period. You must ensure availability for the periods of the 10th May to 1st of July and take particular note of the following deadlines:

• 10th - 15th of May: desk rejection decisions.
• 4th of June: flagging missing/low-quality reviews to PCs.
• 5th of June: contact emergency reviewers (if needed).
• 14th - 21st of June: review consolidation and discussion period.
• 1st of July: submission of the meta-reviews for all papers.

Desk rejections of papers (10th - 15th of May)

Papers for which one of the following is true should be flagged by ACs and desk rejected:

• Is the paper already published in essentially similar form elsewhere?
  1. Prior non-peer-reviewed versions of papers as technical reports or working papers deposited on arXiv or other repositories are acceptable.
  2. A minor overlap (<20%) or an overlap with a paper in a significantly shorter format paper (4 pages or less) is also acceptable.
• Does the paper break anonymity?
  1. For example, by clearly stating the group's name or linking to non-anonymous resources.
• LLMs: Does the paper contain manifestly (and demonstrably) fake references?
• Is the paper not following the format or length?
• Is the paper out of scope (outside of the Computer Vision/AI field)

Additionally, double submission is prohibited, so if it comes to your knowledge or is pointed out by a reviewer that a paper has been submitted for consideration in another venue, the paper should be flagged for desk rejection to the PCs.

Evaluation of received reviews (3rd of June then 10th of June)

At the end of the review period, ACs will do a first evaluation of the received reviews to identify i) missing reviews and ii) low-quality reviews.

A review is considered to be low-quality if the answer to any of the following questions is NO:

• Is the review of sufficient depth?
• Is the review providing a clear motivation for its recommendation?
• Are weaknesses in the paper clearly stated and factual?
• Are claims of lack of novelty specific and supported by references to published (and peer-reviewed) works?
• If a lack of clarity is argued, it should be specific to parts of the paper/approach.
• Similarly, arguments of poor writing should be substantiated with specific examples.

If you identify a low-quality review, you have to query for complement and clarification before the discussion period begins.

Note 1: There will be no rebuttal period for BMVC this year. Therefore reviews should not ask for revisions and/or additional experiments but only evaluate the submitted paper as it stands.

Note 2: The fact that a proposed method does not exceed the state-of-the-art accuracy on an existing benchmark dataset is not grounds for rejection by itself. Instead, it is important to weigh both the novelty and potential impact of the work alongside the reported performance. Minor flaws should not be a reason to reject a paper. Above all, you should be specific and detailed in your reviews.

Discussion period (14th - 21st of June)

Once all reviews have been received and you have queried low-quality reviews for complement and clarification, the discussion period begins. In many instances, all reviewers will agree, and little discussion will be needed, but for some papers, the reviewers will strongly disagree either in their assessment of the paper or, in extreme cases, over factual assessments (e.g. the novelty of the approach or technical soundness).

After the review deadline, the AC of each paper will have to initiate and animate the discussion to reach a consensus and prepare a Meta review. The discussion period aims to answer the following questions:

• After reading the other reviews, did any of the reviewers change their assessment of the paper?
• Where reviewers’ assessments differ, what are the arguments on both sides?
• Where factual disagreements arose, what do both sides provide the evidence?

During the discussion, you should ensure you have all the arguments and facts to reach your decision and write the meta-review.

Making a decision (1st of July)

• You have been selected as an area chair based on your experience and expertise in computer vision and its applications. We trust you to use your best judgment to decide on the papers assigned to you.
• When deciding, please consider the reviewers' comments carefully. In some cases, you may need to look at the paper to form your own opinion, especially in the case of mixed reviews (i.e. when the reviewers' recommendations are not in agreement).
• Overruling all three reviewers to decide on a paper is unacceptable. If you strongly believe in a decision that does not reflect the reviews (e.g., accepting a paper with a majority reject or rejecting a paper with a majority accept), make sure to discuss it with the PCs through email.

Meta-reviews (1st of July)

The meta-reviews are the MOST CRUCIAL aspect of the review process. This is where the Area Chair justifies their recommendation to accept or reject a paper. These meta-reviews should highlight why the decision was reached. If all reviewers agree on a paper, this consolidation report can be simple but feel free to encourage authors and provide constructive feedback. If there is disagreement on the reviews, it is the AC’s job to clarify why you decide to side with either an accept or reject in the face of mixed reviews.

Important guidelines

• The golden rule is to write meta-reviews in a way that you, as an author, would appreciate. No one likes to have their paper rejected, but this experience is even worse when it is done without proper justification and professionalism.
• The goal of the meta-review is primarily to reconcile and adjudicate the reviewers' views. It is not your job to write a fourth review of the paper. Unless absolutely necessary, you should avoid bringing in additional considerations beyond what is found in the reviews. However, in the case of mixed reviews, you can add your opinion of the paper to help break the tie. Be sure to explain carefully in your consolidation report how you came to this decision.
• When arguments from a reviewer were disregarded, a clear argument should be made as to why.
• As mentioned before, it will be rare for you to come to a conclusion that is quite different from that of the three reviewers (e.g. reviews all agree to accept, but you reject the paper). If you feel you need to overrule the reviewers, please bring this to the PCs’ attention.

Note 1: For accepted papers, you should also provide a recommendation to PCs for papers that should be considered for oral presentations rather than posters. Papers that show significant progress to the state of the art and are of broad interest to the BMVC community should be nominated for oral presentations.

Note 2: We would also like ACs to rate their reviewers to allow the allocation of reviewers’ certificates and best reviewers’ awards at the conference. Reviewers should have followed the reviewing guidelines. We encourage you to read this since it may help you assess the review’s quality.

Meta-review templates

You can find below templates that can help you to communicate your decision to the authors.

1. (Rejected Paper)

This paper was reviewed by three experts in the field.  The reviewers raised many concerns regarding the paper, e.g. [RECAP A FEW CONCERNS, SUCH AS NOVELTY, EXPERIMENTAL RESULTS, ETC]. Considering the reviewers’ concerns, we regret that the paper cannot be recommended for acceptance at this time.  The authors are encouraged to consider the reviewers’ comments when revising the paper for submission elsewhere.

2. (Rejected Paper with mixed review) - please acknowledge that the reviews are mixed.

This paper was reviewed by three experts in the field. The reviewers raised the following concerns [RECAP CONCERNS]. Based on the reviews, [give a reason why you side with reviewers’ recommending not to accept the paper - in this case, you are welcome to read the paper and add your opinion as a tiebreaker]. While the paper has merit, the decision is not to recommend acceptance. The authors are encouraged to consider the reviewers’ comments when revising the paper for submission elsewhere.

3. (Accepted Paper with mixed review) - please acknowledge that the reviews are mixed.

This paper was reviewed by three experts in the field. The reviewers raised the following concerns [RECAP CONCERNS]. Based on the reviews, [give a reason why you side with reviewers’ recommending accepting the paper - in this case, you are welcome to read the paper and add your opinion as a tiebreaker]. We congratulate the authors on the acceptance of their paper!

4. (Accept)

Three experts in the field reviewed this paper. Based on their feedback, the decision was made to recommend it for acceptance to BMVC 2024. We congratulate the authors on their acceptance!