entropy based visual attention

import glob
import skimage
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image
import glob

import matplotlib.pyplot as plt
import numpy as np

from skimage.segmentation import felzenszwalb, slic, quickshift, watershed
from skimage.segmentation import mark_boundaries
from skimage.util import img_as_float
from skimage.filters.rank import entropy
from skimage.morphology import disk

from scipy import ndimage as nd
from skimage import feature
from skimage.color import rgb2hsv
import matplotlib.patches as patches

img1_list = glob.glob("*img1.jpg")
img2_list = glob.glob("*img2.jpg")

import PIL

import re

def natural_sort(l): 
    convert = lambda text: int(text) if text.isdigit() else text.lower() 
    alphanum_key = lambda key: [ convert(c) for c in re.split('([0-9]+)', key) ] 
    return sorted(l, key = alphanum_key)

img1s = natural_sort(img1_list)
img2s = natural_sort(img2_list)

for img1_name,img2_name in zip(img1s,img2s):
    pil_image = PIL.Image.open(img1_name)
    pil_image = pil_image.resize((int(pil_image.size[0]/2.0),int(pil_image.size[1]/2.0)), Image.ANTIALIAS)
    rgb = np.array(pil_image)
    gray = np.array(pil_image.convert('L'))
    fig = plt.figure(figsize=(18,9))
    ax1 = fig.add_subplot(1, 2, 1)
    ax2 = fig.add_subplot(1, 2, 2)
    ax1.imshow(rgb)
    print(img1_name)
    #segments_fz = felzenszwalb(rgb, scale=100, sigma=1.5, min_size=150)  #min_size=150, sigma=1.5
    #ax2.imshow(mark_boundaries(rgb, segments_fz))
    #plt.title("Felzenszwalb Segments {}".format(np.unique(segments_fz).shape[0]))
    entropy_img = entropy(gray, disk(25))
    img1 = ax2.imshow(entropy_img, cmap=plt.cm.jet)
    ax2.set_title('Entropy')
    ax2.axis('off')
    fig.colorbar(img1, ax=ax1)
    plt.savefig(img1_name+"_segments.png", dpi=150, transparent=True, bbox_inches='tight')
    plt.close()

    fig = plt.figure(figsize=(18,9))
    ax1 = fig.add_subplot(1,1,1)
    peaks = feature.peak_local_max(entropy_img, min_distance=30, threshold_abs=0.1)
    plt.imshow(rgb)
    plt.plot(peaks[:, 1], peaks[:, 0], 'r.', linewidth=2.0)
    for (y,x) in peaks:
        print(y,x)
        rect = patches.Rectangle((x-32,y-32),64,64,linewidth=1,edgecolor='r',facecolor='none')
        # Add the patch to the Axes
        ax1.add_patch(rect)
    plt.savefig(img1_name+"_candidates.png", dpi=600, transparent=True, bbox_inches='tight')
    plt.close()

    pil_image = PIL.Image.open(img2_name)
    pil_image = pil_image.resize((int(pil_image.size[0]/2.0),int(pil_image.size[1]/2.0)), Image.ANTIALIAS)
    rgb = np.array(pil_image)
    gray = np.array(pil_image.convert('L'))
    fig = plt.figure(figsize=(18,9))
    ax1 = fig.add_subplot(1, 2, 1)
    ax2 = fig.add_subplot(1, 2, 2)
    ax1.imshow(rgb)
    print(img2_name)
    #segments_fz = felzenszwalb(rgb, scale=100, sigma=1.5, min_size=150)  #min_size=150, sigma=1.5
    #ax2.imshow(mark_boundaries(rgb, segments_fz))
    #plt.title("Felzenszwalb Segments {}".format(np.unique(segments_fz).shape[0]))
    entropy_img = entropy(gray, disk(25))
    img1 = ax2.imshow(entropy_img, cmap=plt.cm.jet)
    ax2.set_title('Entropy')
    ax2.axis('off')
    fig.colorbar(img1, ax=ax1)
    plt.savefig(img2_name+"_segments.png", dpi=150, transparent=True, bbox_inches='tight')
    plt.close()

    fig = plt.figure(figsize=(18,9))
    ax1 = fig.add_subplot(1,1,1)
    peaks = feature.peak_local_max(entropy_img, min_distance=30, threshold_abs=0.1)
    plt.imshow(rgb)
    plt.plot(peaks[:, 1], peaks[:, 0], 'r.', linewidth=2.0)
    print(peaks)
    for (y,x) in peaks:
        print(y,x)
        rect = patches.Rectangle((x-32,y-32),64,64,linewidth=1,edgecolor='r',facecolor='none')
        # Add the patch to the Axes
        ax1.add_patch(rect)
    plt.savefig(img2_name+"_candidates.png", dpi=600, transparent=True, bbox_inches='tight')
    plt.close()