Evaluation

The goal of this competition is to determine the catalog ID of stars in several visible scenes. This means primarily to have a robust and correct star identification algorithm and secondarily to have a fast identification as well.

Correctness

The evaluation of correctness of the submission is done automatically and the result will be shown in the leaderboard. Each scene is given a score depending on the number of correctly identified stars:

$$\max\left(\frac{c - 2 w}{\max\left(t, 1\right)}, -1\right),$$ where \(c\) is the number of correctly identified stars, \(w\) is the number of wrongly identified stars and \(t\) is the number of actual stars in the scene. Remember, that a star has a positive identifier and artifacts have the value -1.

The overall score displayed in the leaderboard is the sum over all scene scores.

Speed

The speed of the C source code will be measured with the main.c file given on the scripts page. This process is done manually as soon as possible.

The system used has the following specifications:

• Compiler: gcc (GCC) 5.1.1 20150618 (Red Hat 5.1.1-4)
• CPU: Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz

To be considered in the evaluation, the C code has to produce the same result as the result file in the submission.