Issue with Sparse Dose Distribution Matrix and Convergence in Cervical Cancer Dataset - Clarification on Data Processing and Normalization Steps

[710960345]

Hello, I'm attempting to reproduce your work with my own dataset: the cervical cancer dataset. Unfortunately, I've run into an issue. After calculation, I found that the dose distribution matrix "D" is sparse. This sparsity has caused problems with the convergence within PTV, as can be seen in the figures.

I have several questions :
For data of different shapes or types, I modify the reshape coding. Are there any other crucial parameters here?
After converting the data from the DICOM format to the MAT format (using CERR), is there any additional normalization step involved? If so, is there a defined range for this normalization? In my gantry_couch.mat file, the maximum value of the numerical variable D is 2.5. I'm wondering if this value could have an impact on the convergence of the formula.

I guess the core problem might be related to data format conversion or preprocessing. It would be extremely helpful if you could share more details or relevant code snippets.
Thank you so much for your time and support! Looking forward to your response.

[kels271828]

Hello! What sort of reshaping did you do before running your model? It looks like perhaps a smaller matrix or vector was reshaped into something larger, but you needed to interpolate to fill in missing values?