All reviews follow a double-blind process. All papers are subject to plagiarism analysis using a software tool prior to review.
All regular papers are reviewed by at least two reviewers, but usually by three or more, and rated considering their: Relevance, Originality, Technical Quality, Significance and Presentation;
The reviewers are also asked to answer a group of questions that may help the authors to improve the paper, should it be accepted, namely: Abstract and Introduction are adequate?, Needs more experimental results?, Needs comparative evaluation?, Improve critical discussion?, Figures are Adequate?, Conclusions/Future Work are convincing?, References are up-to-date and appropriate?, Paper formatting needs adjustment?, Improve English?
Finally, the reviewers can provide some free text observations which was given to the authors and also some free text private observations, made available only to the program chair. Conflicting reviews may require assignment of a new reviewer. In the end the program chairs decide. The author has a period for rebuttal, which triggers a workflow involving the chairs and the reviewers if necessary. All rebuttals are answered but decisions are final.