#59: Leverage iterativeMean to not suffer of RAM quantity issue

datasets/raise/split_and_compare_prnus_of_subgroups.py

@@ -2,7 +2,7 @@ from PIL import Image
 import numpy as np
 import matplotlib.pyplot as plt
 from tqdm import tqdm
-from utils import denoise
+from utils import denoise, iterativeMean
 from skimage import img_as_float
 import sys
 import os
@@ -12,7 +12,7 @@ sys.path.insert(0, '../../algorithms/distance/')
 
 from rms_diff import rmsDiffNumpy
 
-NUMBER_OF_SUBGROUPS = 1
+NUMBER_OF_SUBGROUPS = 2
 DENOISER = 'wavelet'
 IMAGES_FOLDER = 'flat-field/TIF'
 
@@ -24,7 +24,8 @@ numberOfImagesThresholds = range(numberOfImagesPerSubgroup, numberOfImagesPerSub
 
 subgroupsPrnuEstimatesNpArray = []
 for subgroupIndex in tqdm(range(NUMBER_OF_SUBGROUPS), 'Subgroup'):
-    imagesPrnuEstimateNpArray = []
+    subgroupPrnuEstimateNpArray = []
+    subgroupIterativeMean = iterativeMean()
     subgroupImagesFileNames = imagesFileNames[numberOfImagesPerSubgroup * subgroupIndex : numberOfImagesPerSubgroup * (subgroupIndex + 1)]
     for imageFileName in tqdm(subgroupImagesFileNames, f'Image of subgroup {subgroupIndex}'):
         imagePath = f'{IMAGES_FOLDER}/{imageFileName}'
@@ -33,11 +34,10 @@ for subgroupIndex in tqdm(range(NUMBER_OF_SUBGROUPS), 'Subgroup'):
         imagePrnuEstimateNpArray = imageNpArray - denoise(imageNpArray, DENOISER)
         imagesPrnuEstimateNpArray += [imagePrnuEstimateNpArray]
 
-    subgroupPrnuEstimateNpArray = []
-    # Not efficient mean computation.
-    for numberOfImagesIndex, numberOfImages in enumerate(numberOfImagesThresholds):
-        subgroupPrnuEstimateNpArray += [np.array(imagesPrnuEstimateNpArray[:numberOfImages]).mean(axis = 0)]
-        plt.imsave('prnu.png', subgroupPrnuEstimateNpArray[-1])
+        subgroupIterativeMean.add(imagePrnuEstimateNpArray)
+        subgroupPrnuEstimateNpArray += [subgroupIterativeMean.mean]
+
+    plt.imsave(f'prnu_subgroup_{subgroupIndex}.png', subgroupPrnuEstimateNpArray[-1])
     subgroupsPrnuEstimatesNpArray += [subgroupPrnuEstimateNpArray]
 
 rmss = []