Merge branch '13-update-eda' into 'main'

Resolve "Update EDA"

Closes #13

See merge request !10

EDA.py

@@ -23,7 +23,7 @@ def find_mean_img(full_mat):
     # calculate the average
         mean_img = np.mean(full_mat[mat], axis = 0)
     # reshape it back to a matrix
-        mean_img = mean_img.reshape((200, 150))
+        mean_img = mean_img.reshape((170, 120))
         ax = fig.add_subplot(rows, cols,i+1)
         ax.imshow(mean_img, vmin=0, vmax=255, cmap='Greys_r')
         ax.set_title('Average ' + mat)
@@ -31,7 +31,7 @@ def find_mean_img(full_mat):
 
     plt.tight_layout()
     
-def plot_pca(pca, title, size = (200, 150)):
+def plot_pca(pca, title, size = (170, 120)):
     """Plots each decomposed PCA image and labels the amount of variability for each image.
 
     Args:
@@ -57,7 +57,7 @@ def plot_pca(pca, title, size = (200, 150)):
     plt.show()
 
 
-def eigenimages(full_mat,n_comp = 0.7, size = (200, 150)):
+def eigenimages(full_mat,n_comp = 0.7, size = (170, 120)):
     """Creates creates and fits a PCA estimator from sklearn.
 
     Args:
@@ -69,6 +69,7 @@ def eigenimages(full_mat,n_comp = 0.7, size = (200, 150)):
         sklearn PCA object: Fitted PCA model.
     """
     # fit PCA to describe n_comp * variability in the class
+    print(full_mat, n_comp)
     pca = PCA(n_components = n_comp, whiten = True)
     pca.fit(full_mat)
 

data_processing.py

@@ -2,7 +2,9 @@ import os
 import shutil
 import pandas as pd
 import tensorflow as tf
+from PIL import Image
 from tensorflow.keras.preprocessing import image
+from tensorflow.keras.layers import CenterCrop
 import numpy as np
 from tqdm import tqdm
 
@@ -36,16 +38,23 @@ def load_sort_data(meta_filename = str, image_folder = str, output_folder = str)
     
     return metadata, dest_dir
 
-def transform(path, size = (300, 225)):
+def transform(path, size = (200, 150)):
     # create a list of images
     img_list = [fn for fn in os.listdir(path) if fn.endswith('.jpg')]
+    (left, upper, right, lower) = (15, 15, 135,185)
     #iterating over each .jpg
     for fn in tqdm(img_list):
         fp = path + '/' + fn
         current_image = image.load_img(fp, target_size = size, 
                                     color_mode = 'grayscale')
-        # covert image to a matrix
-        img_ts = image.img_to_array(current_image)
+        crop_image = current_image.crop((left, upper, right, lower))
+        
+        # print("current image", type(current_image), current_image)
+        # print("crop image", type(crop_image), crop_image)
+        # convert image to a matrix
+        img_ts = image.img_to_array(crop_image)
+        
+        # print(type(img_ts), img_ts)
         # turn that into a vector / 1D array
         img_ts = [img_ts.ravel()]
         try: