F.A.Q.¶

This page contains the answers to Frequently Asked Question(s) (F.A.Q.) and will be gradually populated.

Q1: It appears that data starting with "P" seems to be missing information related to the upper part of the oral cavity, including the upper teeth, maxilla, maxillary sinus, etc. Could you please confirm if this is indeed the case for the dataset?

A2: Yes, that is correct: handling with such a different view is part of the challenge.  The dataset can be divided in "Set A" (samples that start with P) and "Set B" (patients that begin with "F") . For more details, please refer to Ditto. Test data are assured to have the same field of view as "F" cases.

Q2: As mentioned on the dataset homepage, the "Dataset classes" are listed as 42. However, in the dataset.json file, I find a total of 48 classes.

A2: Classes are 42, as specified on our webpage. Numbering in the dataset.json file goes up to 48, but if you look at the classes, you'll see that six of them are Not Available ("NA", i.e., 19, 20, 29, 30, 39, 40). This is just to keep the teeth identifiers aligned with the standard medical notation. Thus, only 42 classes should be predicted by your algorithm. 

Q3:   Is the direction information stored in the .mha files physically accurate? I have observed that all images in the dataset are in the 'RAI' direction. Will there be images with a different direction in the Final Test Phase? If so, will the direction information be stored accurately in the file metadata?

A3: Yes, the metadata is physically accurate and will also be so in the test set. Anyway, all images selected for the Final Test Phase will have RAI direction.