Dear SMILE-UHURA Participants, We are pleased to announce that the Training-Validation dataset for our challenge has been made public. Go to the Dataset section on Synapse for the same (https://www.synapse.org/#!Synapse:syn47164761/datasets/) We have also split the Training-Validation dataset into training and validation subsets, which can be found in the Files section (https://www.synapse.org/#!Synapse:syn47164761/files/) This split is just for reference purposes. You are free to use any split you like, including training on the whole dataset. Folder structure (as per Dataset section) [in brackets, as per Files section] **TrainVal Volumes** [_TrainingSet/train_ and _ValidationSet/validate_] : Actual input volumes from the [StudyForrest](https://www.studyforrest.org/) dataset **TrainVal Manual Labels** [_TrainingSet/train_label_ and _ValidationSet/validate_label_] : Manually annotated labels, to be used for training and evaluation. **TrainVal OMELETTE Labels** [_TrainingSet/train_OMELETTElabel _and _ValidationSet/validate_OMELETTElabel_] : Additional labels, created in an automated fashion using the [OMELETTE](https://gitlab.com/hmattern/omelette) pipeline - using two different sets of parameters (i.e. two different labels for each volume are available). These labels can be used as additional training labels, if desired, to have multiple labels for each volume during training. **TrainVal Plausible Labels** [_TrainingSet/train_plausiblelabel_ and _ValidationSet/validate_plausiblelabel_] : 10 plausible segmentations for each of the volumes were created in a semi-automatic fashion by manipulating the parameters of the Frangi filter. These can also be used as additional training labels, if desired. Using these labels, methods like Probabilistic UNet can be trained. There will be two different test datasets - public and private. The public test set will be released (without any labels, only the actual volumes) on 15.02.2023. The private test set will never be made public. Once the dockers are submitted, separate evaluations will be performed on the public, and private test sets and two sets of results will be released. Feel free to use this discussion thread to talk about the dataset - to post your questions, confusions, suggestions, anything related to the Dataset. Looking forward to your submissions :) With Regards, Organising Team SMILE-UHURA Challenge