While running the 10k series cohort, 6 series failed at one step. This was different from the first iteration. Upon request to save the raw radiomics features (which will include general features in addition to first and shape features), I added a step for this purpose. I have now also added an error-handling step to tackle such issues.
Additionally, while generating manifests for the 1035 and 10k series cohorts, I noticed that the CPU and RAM assignments for the DICOM segmentation and radiomics steps were lost in previous PRs. I have since added them back.
While running the 10k series cohort, 6 series failed at one step. This was different from the first iteration. Upon request to save the raw radiomics features (which will include general features in addition to first and shape features), I added a step for this purpose. I have now also added an error-handling step to tackle such issues.
Additionally, while generating manifests for the 1035 and 10k series cohorts, I noticed that the CPU and RAM assignments for the DICOM segmentation and radiomics steps were lost in previous PRs. I have since added them back.