Ranking¶

We will provide both an overall ranking by averaging the individual rankings of the scores of each metric used.

Anyone can take part in the challenge. Participants will be displayed in order of scores, and winners will be determined by the highest overall average on all individuals rankings.

Scores will evaluate all participants, regardless of code type and model weight and availability, so as not to limit participation by companies even if they do not wish to distribute their code for intellectual property or commercial reasons (but cannot be on the final published winners).

Bonuses¶

In addition, participants can earn bonus points for aspects such as:

• code Documentation
• code Reusability
• code Accessibility
• and code quality  : Codacy
• novelty (metadata-aware training strategies, etc.)

Of course, the bonuses will have no impact on the final ranking, but the methods with the most bonuses will have their place and will be described in the article.

Metrics¶

We will use standard metrics that are used in the field to evaluate the submissions: More details coming soon ; we will use at least :

• Structural Similarity Index Measure (SSIM) of predicted and ground truth images.

• Pearson Correlation Coefficient (PCC) of predicted and ground truth images.

NB : All cases are not always filled.

As some test images do not contain all the organelles, the metrics will only be calculated if the ground truth image exists, but it is important to predict the 4 output images.