Regarding Distinct Lesion ¶
By: ghanshyamd on June 3, 2025, 8:57 a.m.
Hello Organizers, I have following queries regarding the data. Query 1: After performing connected component analysis on the given labeled images, I have found that quite a lot of cases have multiple lesions with size less than 5 voxels. Are these lesions considered as noise or should they be included in the analysis? If they will be removed what will be the criteria for removal?
Query 2: Since we are supposed to perform binary segmentation on the images, some of the label images have multiple label values (ex: Patient 0894). Currently I am moving forward by merging these labels into a single label. Hope this assumption is correct. If not, please clarify further.
