Odtnhj

I want to create a list of 10 matrix (32,32,3) that will contain a sum of values from my RGB images.

So I used matrixSum = np.zeros((10,32,32,3),np.uint64)

And I thought to access it using matrixSum[indiceClass][ligne][col][rgb]

But I keep having an error telling me that matrixSum[indiceClass] is only size 1


So I changed my code to matrixSum[indiceClass][0][ligne][col][rgb] and I do not have this error anymore but it brings me troubles later.

Am I doing something wrong? It don't understand why I would have an array of size 1 in the middle instead of 32. Or is it logical?

Here is the full code of the function : (indiceclasse will only be between 0 and 9)

def createMatrixSum(data, nbImagesParClasse):
 matricesSum = np.zeros((10,32,32,3),np.uint64)
 indiceClasse = 0
 for i in range(0,80):
 indiceClasse = data["y"][i]
 indiceClasse -= 1;
 nbImagesParClasse[indiceClasse] += 1
 for ligne in range(32):
 for col in range(32):
 for rgb in range(3):
 matricesSum[indiceClasse][ligne][col][rgb] += data["X"][ligne,col,rgb,i]
 return matricesSum, nbImagesParClasse

Edit : here is the len and shape of my data/matrix before the error.


I don't see anything wrong.

Well, turns out that the "indiceClass" I tried to obtain was not an integer as I thought but a list of size 1.
So using it as an index added another list into my matrix which explain why I got that list into the middle. Should have checked that beforehand.