The RSNA has officially launched its Intracranial Hemorrhage Detection and Classification artificial intelligence (AI) challenge.
The RSNA's third annual AI challenge will test the ability of AI algorithms to identify and classify subtypes of hemorrhages on head CT scans. The challenge dataset -- the first multiplanar dataset used in an RSNA AI challenge -- includes more than 25,000 head CT scans contributed by several research institutions. All studies have been labeled for the presence of five subtypes of intracranial hemorrhage by volunteer specialists from the American Society of Neuroradiology, according to the RSNA.
The society said the first wave of data was released on September 3 to researchers for use in training and developing algorithms. Participants will then apply their algorithms to the testing portion of the dataset from November 4 to 11. After these results are compared with the annotations on the testing dataset, an evaluation metric will be used to rate the algorithms' accuracy and determine the winners. The results will be announced in November; the top submissions will be recognized at the AI Showcase Theater during RSNA 2019 in Chicago in December.
The challenge is being run on a platform provided by Kaggle, who has recognized it as a public good and will award $25,000 to the winning entries, according to the society. More information on the challenge is available on the RSNA's website.