sid321axn

n_samples = 5  # number of samples of each cancer type

# Plotting code
fig, m_axs = plt.subplots(7, n_samples, figsize = (4*n_samples, 3*7))
for n_axs, (type_name, type_rows) in zip(m_axs, 
                                         skin_df.sort_values(['cell_type']).groupby('cell_type')):
    n_axs[0].set_title(type_name)
    for c_ax, (_, c_row) in zip(n_axs, type_rows.sample(n_samples, random_state=1234).iterrows()):
        c_ax.imshow(c_row['image'])
        c_ax.axis('off')

sid321axn

skin_df = pd.read_csv(os.path.join(base_skin_dir, 'HAM10000_metadata.csv'))

# Creating New Columns for better readability

skin_df['path'] = skin_df['image_id'].map(imageid_path_dict.get)
skin_df['cell_type'] = skin_df['dx'].map(lesion_type_dict.get) 
skin_df['cell_type_idx'] = pd.Categorical(skin_df['cell_type']).codes

# Now lets see the sample of tile_df to look on newly made columns
skin_df.head()

sid321axn

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import os
from glob import glob
import seaborn as sns
from PIL import Image
np.random.seed(123)
from sklearn.preprocessing import label_binarize

sid321axn

base_skin_dir = os.path.join('..', 'input')

# Merging images from both folders HAM10000_images_part1.zip and HAM10000_images_part2.zip into one dictionary

imageid_path_dict = {os.path.splitext(os.path.basename(x))[0]: x
                     for x in glob(os.path.join(base_skin_dir, '*', '*.jpg'))}

# This dictionary is useful for displaying more human-friendly labels later on

lesion_type_dict = {

sid321axn

features=skin_df.drop(columns=['cell_type_idx'],axis=1)

target=skin_df['cell_type_idx']

sid321axn

x_train_o, x_test_o, y_train_o, y_test_o = train_test_split(features, target, test_size=0.20,random_state=123)

sid321axn

x_train = np.asarray(x_train_o['image'].tolist())
x_test = np.asarray(x_test_o['image'].tolist())

x_train_mean = np.mean(x_train)
x_train_std = np.std(x_train)

x_test_mean = np.mean(x_test)
x_test_std = np.std(x_test)

x_train = (x_train - x_train_mean)/x_train_std

sid321axn

# Perform one-hot encoding on the labels
y_train = to_categorical(y_train_o, num_classes = 7)
y_test = to_categorical(y_test_o, num_classes = 7)

sid321axn

x_train, x_validate, y_train, y_validate = train_test_split(x_train, y_train, test_size = 0.1, random_state = 2)

sid321axn

# Reshape image in 3 dimensions (height = 75px, width = 100px , canal = 3)
x_train = x_train.reshape(x_train.shape[0], *(75, 100, 3))
x_test = x_test.reshape(x_test.shape[0], *(75, 100, 3))
x_validate = x_validate.reshape(x_validate.shape[0], *(75, 100, 3))