diff --git a/mdata/ILSVRC.py b/mdata/ILSVRC.py
index c1dfa1f..ed04903 100644
--- a/mdata/ILSVRC.py
+++ b/mdata/ILSVRC.py
@@ -33,7 +33,8 @@ package_dir = os.path.dirname(os.path.abspath(__file__))
class DataILSVRC(DataDumperBase):
- def __init__(self, base_dir='/media/chunk/Elements/D/data/ImageNet/img/ILSVRC2013_DET_val', category='Train'):
+ def __init__(self, base_dir='/media/chunk/Elements/D/data/ImageNet/img/ILSVRC2013_DET_val', category='Train',
+ host='HPC-server'):
DataDumperBase.__init__(self, base_dir, category)
self.base_dir = base_dir
@@ -45,6 +46,10 @@ class DataILSVRC(DataDumperBase):
self.feat_dir = os.path.join(self.dst_dir, 'Feat')
self.img_dir = os.path.join(self.dst_dir, 'Img')
+ self.host = host
+ self.master = 'spark://%s:7077' % self.host
+ self.appname = 'ImageILSVRC'
+
self.dict_data = {}
self.table_name = self.base_dir.strip('/').split('/')[-1] + '-' + self.category
@@ -316,7 +321,7 @@ class DataILSVRC(DataDumperBase):
return self.table
if self.connection is None:
- c = happybase.Connection('HPC-server')
+ c = happybase.Connection(host=self.host)
self.connection = c
tables = self.connection.tables()
@@ -345,7 +350,7 @@ class DataILSVRC(DataDumperBase):
table_name = self.table_name
if self.connection is None:
- c = happybase.Connection('HPC-server')
+ c = happybase.Connection(host=self.host)
self.connection = c
tables = self.connection.tables()
@@ -506,7 +511,7 @@ class DataILSVRC(DataDumperBase):
elif mode == "spark": # cluster
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageCV', master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname, master=self.master)
result = self.sparker.read_hbase(self.table_name) # result = {key:[feat,tag],...}
for feat, tag in result:
diff --git a/mdata/ILSVRC_S.py b/mdata/ILSVRC_S.py
index e98089a..1734990 100644
--- a/mdata/ILSVRC_S.py
+++ b/mdata/ILSVRC_S.py
@@ -39,7 +39,7 @@ class DataILSVRC_S(DataDumperBase):
copyright(c) 2015 chunkplus@gmail.com
"""
- def __init__(self, base='ILSVRC2013_DET_val', category='Train_1', tablename=None):
+ def __init__(self, base='ILSVRC2013_DET_val', category='Train_1', host='HPC-server', tablename=None):
DataDumperBase.__init__(self, base, category)
self.base = base
@@ -55,8 +55,11 @@ class DataILSVRC_S(DataDumperBase):
else:
self.table_name = tablename
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.host = host
+ self.master = 'spark://%s:7077' % self.host
+ self.appname = 'ImageILSVRC-S'
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
self.steger = F5.F5(sample_key, 1)
@@ -67,7 +70,7 @@ class DataILSVRC_S(DataDumperBase):
return self.table
if self.connection is None:
- c = happybase.Connection('HPC-server')
+ c = happybase.Connection(host=self.host)
self.connection = c
tables = self.connection.tables()
@@ -91,7 +94,7 @@ class DataILSVRC_S(DataDumperBase):
table_name = self.table_name
if self.connection is None:
- c = happybase.Connection('HPC-server')
+ c = happybase.Connection(host=self.host)
self.connection = c
tables = self.connection.tables()
@@ -251,8 +254,8 @@ class DataILSVRC_S(DataDumperBase):
elif mode == 'spark':
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
cols = [
'cf_pic:data',
@@ -285,8 +288,8 @@ class DataILSVRC_S(DataDumperBase):
withdata=withdata)
elif mode == 'analysis':
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
cols = [
'cf_pic:data',
@@ -401,8 +404,8 @@ class DataILSVRC_S(DataDumperBase):
elif mode == 'spark':
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
cols = [
'cf_pic:data',
@@ -496,8 +499,8 @@ class DataILSVRC_S(DataDumperBase):
elif mode == 'spark':
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
cols = [
'cf_pic:data',
@@ -532,8 +535,8 @@ class DataILSVRC_S(DataDumperBase):
def _analysis(self, mode='analysis', feattype='ibd', readforward=False, writeback=True, withdata=False):
if mode == 'analysis':
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
cols = [
'cf_pic:data',
@@ -618,8 +621,8 @@ class DataILSVRC_S(DataDumperBase):
elif mode == "spark" or mode == "cluster":
if self.sparker == None:
- self.sparker = SC.Sparker(host='HPC-server', appname='ImageILSVRC-S',
- master='spark://HPC-server:7077')
+ self.sparker = SC.Sparker(host=self.host, appname=self.appname,
+ master=self.master)
rdd_dataset = self.sparker.read_hbase(self.table_name, func=rdd.rddparse_dataset_ILS, collect=False)
if not collect:
diff --git a/msteg/steganalysis/ChiSquare.py b/msteg/steganalysis/ChiSquare.py
deleted file mode 100644
index 7d0bb2c..0000000
--- a/msteg/steganalysis/ChiSquare.py
+++ /dev/null
@@ -1,162 +0,0 @@
-"""
-
-This module implements an algorithm described by Andreas Westfeld in [1,2],
-which detects if there was data embedded into an image using JSteg.
-It uses the property that JSteg generates pairs of values in the
-DCT-coefficients histogram, which can be detected by a \chi^2 test.
-
-
-
-[1]: Andreas Westfeld, F5 - A Steganographic Algorithm High Capacity Despite
-Better Steganalysis
-[2]: Andreas Westfeld, Angriffe auf steganographische Systeme
-
-"""
-
-from collections import defaultdict
-import os
-
-from PIL import Image
-import numpy
-from scipy.stats import chisquare
-import matplotlib.pyplot as plt
-import itertools as it
-
-from .. import *
-
-
-class ChiSquare(StegBase):
- """
- The module contains only one method, detect.
- """
-
- def __init__(self, ui, core):
- self.ui = ui
- self.core = core
-
- def detect(self, src, tgt, tgt2):
- """
-
- Detect if there was data embedded in the source image image with
- JSteg algorithm.
-
-
-
- Parameters:
-
- Source image
Image which should be testedTarget image
Image which displays a graphic with the
- embedding probability2nd Target image
Image which displays the embedding
- positions in the image
-
- """
- # --------------------------- Input -----------------------------------
- # If src is from the image pool, test whether the image exists encoded
- # on the file system. Otherwise we can not read DCT-coefficients.
- if self.core.media_manager.is_media_key(src):
- src = self.core.media_manager.get_file(src)
- if hasattr(src, 'tmp_file'):
- src = src.tmp_file
- self.ui.display_error('Trying file: %s' % src)
- else:
- self.ui.display_error('Can not detect anything from \
- decoded images.')
- return
- # Test whether the file exists.
- if not os.path.isfile(src):
- self.ui.display_error('No such file.')
- return
- # Test if it is a JPEG file.
- if not self._looks_like_jpeg(src):
- self.ui.display_error('Input is probably not a JPEG file.')
- return
-
- # ---------------------------- Algorithm ------------------------------
- # Build DCT-histogram in steps of \approx 1% of all coefficients and
- # calculate the p-value at each step.
-
- # dct_data = rw_dct.read_dct_coefficients(src)
- dct_data = self._get_cov_data(src)
-
- hist = defaultdict(int)
- cnt = 0
- l = len(dct_data)
- one_p = l / 100
- result = []
- for block in dct_data:
- # update the histogram with one block of 64 coefficients
- for c in block:
- hist[c] += 1
-
- cnt += 1
- if not cnt % one_p:
- # calculate p-value
- self.ui.set_progress(cnt * 100 / l)
-
- # ignore the pair (0, 1), since JSteg does not embed data there
- hl = [hist[i] for i in range(-2048, 2049) if not i in (0, 1)]
- k = len(hl) / 2
- observed = []
- expected = []
- # calculate observed and expected distribution
- for i in range(k):
- t = hl[2 * i] + hl[2 * i + 1]
- if t > 3:
- observed.append(hl[2 * i])
- expected.append(t / 2)
- # calculate (\chi^2, p)
- p = chisquare(numpy.array(observed), numpy.array(expected))[1]
- result.append(p)
-
- # ----------------------------- Output --------------------------------
- # Graph displaying the embedding probabilities in relation to the
- # sample size.
- figure = plt.figure()
- plot = figure.add_subplot(111)
- plot.grid(True)
- plot.plot(result, color='r', linewidth=2.0)
- plt.axis([0, 100, 0, 1.1])
- plt.title('Embedding probability for different percentages \
-of the file capacity.')
- plt.xlabel('% of file capacity')
- plt.ylabel('Embedding probability')
-
- if self.core.media_manager.is_media_key(tgt):
- img = figure_to_pil(figure)
- self.core.media_manager.put_media(tgt, img)
- else:
- plt.savefig(tgt)
-
- # Image displaying the length and position of the embedded data
- # within the image
- img2 = Image.open(src)
- img2.convert("RGB")
- width, height = img2.size
-
- for i in range(100):
- result[i] = max(result[i:])
-
- cnt2 = 0
- for (top, left) in it.product(range(0, height, 8), range(0, width, 8)):
- if not cnt2 % one_p:
- r = result[cnt2 / one_p]
- if r >= 0.5:
- color = (255, int((1 - r) * 2 * 255), 0)
- else:
- color = (int(r * 2 * 255), 255, 0)
- cnt2 += 1
- img2.paste(color, (left, top, min(left + 8, width),
- min(top + 8, height)))
- self.core.media_manager.put_media(tgt2, img2)
-
- def __str__(self):
- return 'Chi-Square-Test'
-
-
-def figure_to_pil(figure):
- figure.canvas.draw()
- return Image.fromstring('RGB',
- figure.canvas.get_width_height(),
- figure.canvas.tostring_rgb())
diff --git a/test/test_jpeg.py b/test/test_jpeg.py
index e3aa247..fbbaf60 100644
--- a/test/test_jpeg.py
+++ b/test/test_jpeg.py
@@ -1,8 +1,6 @@
__author__ = 'chunk'
import numpy as np
-import matplotlib.pyplot as plt
-import seaborn as sns
from .. import mjpeg
from ..mjpeg import base
from ..msteg.steganography import LSB, F3, F4, F5
@@ -24,9 +22,6 @@ sample_key = [46812L, 20559L, 31360L, 16681L, 27536L, 39553L, 5427L, 63029L, 565
61908L, 63014L,
5908L, 59816L, 56765L]
-# plt.ticklabel_format(style='sci', axis='both', scilimits=(1, 4))
-
-plt.ticklabel_format(style='sci', axis='both')
package_dir = os.path.dirname(os.path.abspath(__file__))
@@ -182,38 +177,6 @@ def test_jpeg():
print mjpeg.diffblocks(ima, imc)
-def test_hist():
- ima = mjpeg.Jpeg(os.path.join(package_dir, "../res/high/pic3_orig.jpg"), key=sample_key)
- print ima.getQuality()
- print ima.getCapacity('All')
-
- capacity = ima.getCapacity()
- print capacity
- rate = 0.65
- hidden = np.random.bytes(int(int(capacity) * rate) / 8)
- steger = F5.F5(sample_key, 1)
- steger2 = F4.F4(key=None)
- steger3 = LSB.LSB(key=None)
- embed_rate = steger3.embed_raw_data(os.path.join(package_dir, "../res/high/pic3_orig.jpg"),
- hidden,
- os.path.join(package_dir, "../res/high/pic3_dest.jpg"),
- frommem=True)
- #
- print embed_rate
-
- imb = mjpeg.Jpeg(os.path.join(package_dir, "../res/high/pic3_dest.jpg"), key=sample_key)
- print imb.getQuality()
- print imb.getCapacity('All')
-
- A = imb.rawsignal().tolist()
- E = [i for i in range(-8, 10)]
- plt.hist(A, E, histtype='bar', rwidth=0.8, align='left')
- plt.xlabel("JPEG coefficients after quantisation")
- plt.ylabel("Frequency")
- plt.xticks([i for i in range(-8, 9)])
- plt.ylim(ymax=300000)
- plt.show()
-
if __name__ == '__main__':
# timer.mark()
--
libgit2 0.21.2