Image_Process_KMeans.py

# -*- coding: utf-8 -*-
"""Untitled0.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/14LTyfo6C3YR0y5p0GFNNPJ1XuDnw0Upb
"""

# -*- coding: utf-8 -*-
"""Untitled4.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1wGSJw9VOMvkf7neop8LraseAL48XC_Ov
"""

from google.colab import drive
drive.mount('/content/drive/')
#https://drive.google.com/open?id=1z-PTDuxVbVUZzbr080VZrSrdL5gqiYeL

!ls "/content/drive/My Drive/Datasets"

from keras.preprocessing import image
from keras.applications.vgg16 import VGG16
from keras.applications.vgg16 import preprocess_input
import numpy as np
from sklearn.cluster import KMeans
import os, shutil, glob, os.path
from PIL import Image as pil_image
image.LOAD_TRUNCATED_IMAGES = True 
model = VGG16(weights='imagenet', include_top=False)

# Variables
imdir = '/content/drive/My Drive/Datasets/class-3_pearlite'

number_clusters = 3

# Loop over files and get features
filelist = glob.glob(os.path.join(imdir, '*.tif'))
filelist.sort()
featurelist = []

for i in range(0,len(filelist)):
  #print(i)
  imagepath=filelist[i]
  img = image.load_img(imagepath, target_size=(224, 224))
  img_data = image.img_to_array(img)
  img_data = np.expand_dims(img_data, axis=0)
  img_data = preprocess_input(img_data)
  features = np.array(model.predict(img_data))
  featurelist.append(features.flatten())

# Clustering
kmeans = KMeans(n_clusters=number_clusters, random_state=0).fit(featurelist)

#print(kmeans)
import pandas as pd
df = pd.DataFrame(columns=['sample', 'output_cluster'])
for i in range(0,len(filelist)):
  m=kmeans.labels_[i]
  df1={'sample':filelist[i].rpartition('/')[-1],'output_cluster':m}
  df = df.append(df1,ignore_index=True)
print(df)


writer = pd.ExcelWriter('output.xlsx')
df.to_excel(writer)
writer.save()
from google.colab import files
files.download('output.xlsx')