import numpy from scipy import misc from watermarker import Watermarker from test import test_jpg, add_blocks, test_filter from skimage.exposure import rescale_intensity w = Watermarker(6, 4) w2 = Watermarker(6, 4, mother="haar", seed=212219812811) # src = misc.imread("../qrmark/pics/sky.png") src = misc.imread("pics/face.jpg") # src = misc.lena() out = w.embed(src, "123456", k=3, tv_denoising_weight=0) out = w2.embed(out, "789012", k=4, tv_denoising_weight=0) misc.imsave("orig.png", out) out2 = misc.imread("orig.png") print w.extract(out2) print "min jpg quality:", test_jpg(w, out2) print w2.extract(out2) print "min jpg quality:", test_jpg(w2, out2) # import random # random.seed(3286912) # print "Max random block coverage: ", test_filter(w, out2, "blocks-%s.png", # lambda img, k: add_blocks(img, k, 50, 50), [i/20.0 for i in range(1, 20)])
import os import numpy from scipy import misc from watermarker import Watermarker from test import test_jpg, add_blocks, add_noise, test_filter, test_recursive_filter from scipy.ndimage.filters import gaussian_filter, gaussian_laplace, uniform_filter, median_filter from skimage.filter import tv_denoise w = Watermarker(6, 4) src = misc.imread("scans/facescan.png") print w.extract(src) print test_jpg(w, src) #src = misc.lena() #markers = [Watermarker(6, 4, seed = seed) for seed in range(50)] #text = "the quick brown fox jumped over the lazy dog and laughed a great deal" # #for i, w in enumerate(markers): # src = w.embed(src, text[i:i+6]) # print "%s\t%s" % (i+1, test_jpg(markers[0], src)) #out = w2.embed(out, "789012", k = 4, tv_denoising_weight = 0) #out2 = misc.imread("orig.png") #out2 = misc.imread("facescan.png") #print w.extract(out2) #print "min jpg quality:", test_jpg(w, out2) #print w2.extract(out2) #print "min jpg quality:", test_jpg(w2, out2)
import os import numpy from scipy import misc from watermarker import Watermarker from test import test_jpg, add_blocks, add_noise, test_filter, test_recursive_filter from scipy.ndimage.filters import gaussian_filter, gaussian_laplace, uniform_filter, median_filter from skimage.filter import tv_denoise src = misc.lena() markers = [Watermarker(6, 4, seed = seed) for seed in range(50)] text = "the quick brown fox jumped over the lazy dog and laughed a great deal" for i, w in enumerate(markers): src = w.embed(src, text[i:i+6]) print "%s\t%s" % (i+1, test_jpg(markers[0], src)) #out = w2.embed(out, "789012", k = 4, tv_denoising_weight = 0) #out2 = misc.imread("orig.png") #out2 = misc.imread("facescan.png") #print w.extract(out2) #print "min jpg quality:", test_jpg(w, out2) #print w2.extract(out2) #print "min jpg quality:", test_jpg(w2, out2) #import random #random.seed(3286912) #print "Max random block coverage: ", test_filter(w, out2, "blocks-%s.png", # lambda img, k: add_blocks(img, k, 50, 50), [i/20.0 for i in range(1, 20)])