CVPR 2027 · scores and rebuttal
What do your CVPR review scores mean?
Start with each reviewer’s recommendation and the reasons behind it. Use those reasons to plan your rebuttal; a mean score alone cannot tell you whether your paper will be accepted.
Last checked against official sources on .
CVPR 2027 review and rebuttal dates
- Reviews released: January 25, 2027.
- Author rebuttals due: February 1, 2027.
- Final decisions: February 25, 2027.
Deadlines use Anywhere on Earth. Check the official schedule for updates.
Before reviews arrive, use the CVPR review-form explainer to understand what reviewers assess.
Read the recommendation label first
- Strong Accept
- A strong positive recommendation. Check whether other reviewers identify a substantive issue that this review does not address.
- Weak Accept
- A positive recommendation with reservations. Identify which reservations you can resolve with evidence from the paper.
- Borderline
- An undecided recommendation. Look for the specific question or missing clarification that could change the assessment.
- Weak Reject
- A negative recommendation. Address the main reason for rejection before spending space on minor corrections.
- Reject
- A negative recommendation that needs a substantive answer. Distinguish a factual misunderstanding from a limitation your paper actually has.
These interpretations are practical guidance for authors, not official acceptance thresholds.
Can CVPR scores tell you your acceptance odds?
We do not have a verified, representative dataset pairing CVPR review scores with both accepted and rejected outcomes, so we do not publish a score-to-acceptance percentage. A conference’s overall acceptance rate does not give the probability for your particular set of reviews.
Averaging numbers also hides disagreement and the reasons behind it. Two positive reviews and one negative review can mean a correctable misunderstanding, or a technical problem the other reviewers missed. The 2027 reviewer guidelines emphasize detailed arguments over the score; area chairs combine the assessments into a recommendation.
A simulated review is practice feedback. Its scores are not calibrated CVPR acceptance probabilities.
How to read mixed or borderline reviews
- Mostly positive: check whether any unresolved criticism challenges a central claim. Do not assume the positive count settles it.
- Borderline recommendations: identify what evidence or clarification each reviewer says would change their assessment.
- A split panel: compare the reasons. Prioritize correctness, evaluation and novelty concerns over trying to persuade reviewers with the average score.
- Mostly negative: separate misunderstandings you can answer from substantive limitations. Be explicit about what the submission establishes and what it does not.
Use the week to answer the concerns that matter
- List each reviewer’s main concern and the evidence needed to answer it. Group repeated concerns together.
- Draft a direct answer, then cite the relevant section, table or figure in your submission.
- Give the most space to issues that could change the recommendation. Keep acknowledgements and minor corrections brief.
- Check that every claim in the response is supported, then fit it into the official one-page template.
Under the 2027 author rules, the rebuttal must remain anonymous and contain no external links. Avoid new contributions or experimental results unless specifically requested by a reviewer. The CVPR rebuttal guide covers the format and includes a copyable skeleton.
Turn your CVPR reviews into a response plan
Paste reviews of your own paper, then attach it for draft replies that cite its evidence. Importing is free; drafting comes with a paper unlock or Pro.
Sources
reviewpaper.ai is not affiliated with CVPR, the CVF or IEEE or its organizers. Rules change between cycles; the official pages above always win.