def main(imgname):
global mycompress,myoutfile
from FFT2util import FFT2D,iFFT2D,matslice,realintmat
from FFTjan import mat2dict,dict2mat
from Tools import find_cutoff_nd
data, dim = img3mat(imgname)
print "starting encoding of image %s" % (imgname)
encoded = map(FFT2D,data)
print "encoding succesfull"
dim2 = (len(encoded[0][0]),len(encoded[0]))
print dim2
import numpy as np
cutoff = find_cutoff_nd( np.array(encoded), mycompress )
print "making dicks"
dicks = map(lambda x: mat2dict(x,cutoff),encoded)
print len(dicks[0])
print map(lambda x: compress(x[0],x[1]), zip(dicks,encoded))
print "listing dicks"
lists = map(lambda x: dict2mat(x,dim2[0],dim2[1]),dicks)
print "starting decoding"
decoded = map(iFFT2D,lists)
print "decoding succesfull"
sliced = map(lambda x: matslice(realintmat(x),dim),decoded)
if myoutfile != None:
filename = myoutfile
else:
filename = imgname.replace('.','_new.')
print "saving to",filename
mat3img(sliced,dim).save(filename)
def main():
x = np.array(
[[[1,1,1,1],
[1,1,1,1],
[1,1,0,1],
[1,1,1,1]],
[[1,1,1,1],
[1,1,1,1],
[1,1,1,1],
[1,1,1,1]],
[[1,1,1,1],
[1,1,1,1],
[1,1,2,1],
[1,1,1,1]],
[[1,1,1,1],
[1,1,1,1],
[1,1,3,1],
[1,1,1,1]]]
)
x_float = x.astype('float')
wavelet = HaarWavelet
X = wavelet.dwt(x_float)
cutoff = find_cutoff_nd( X, 0.3 )
print X
print cutoff
dict = mat_3d_to_dict( X, cutoff )
X_ = dict_to_mat_3d( dict, X.shape[0], X.shape[1], X.shape[2] )
print dict
x_ = wavelet.idwt(X_)
print x_
print PSNR( x_float, x_ )
data = map(lambda x: np.array(x), data)
data_float = map(lambda x: x.astype('float'),data)
print "Starting encoding of image %s" % (sys.argv[1])
print "Image has dimensions : %i,%i" % dim
print
encoded = map(mywavelet.dwt,data_float)
interdims = encoded[0].shape
print "Encoding done"
print "Using dimensions:",interdims
print
# Temporary bypass to check for problems in main algorithm -- Confirmed
compression = find_cutoff_nd( encoded, 0.0 )
print "Converted to dictionaries:"
print "Compression cutoff is %f" % compression
dicks = map(lambda x: mat_3d_to_dict(x,compression),encoded)
print "Dictionary lengths:"
for x in dicks:
print len(x)
print "Compression factor per channel:"
print map(lambda x: compress(x[0],x[1]), zip(dicks,encoded))
print
print "Converting dictionaries to matrices"
def plaatje(plaat):
global mywavelet, mycompress, myoutfile, mytensor, mydelay, mytensor, mythreshfunc
print "compressing",plaat
print "tensor is %s" % ("on" if mytensor else "off")
print "thresholding :",mythreshfunc
print "splitting up gif... (this may take a while)"
call(["gifsicle", "--unoptimize", "-e", plaat, "-o", plaat])
gifs = sorted(glob(''.join([plaat,".*"])))
driedlist = [[],[],[],[]]
for f in gifs:
data, dim = img3mat( f )
for i in range( len(data) ):
driedlist[i].append(data[i])
data = map( lambda x: np.array( x ), driedlist )
print np.array(data).shape
data_float = map( lambda x: x.astype('float'), data )
encoded = np.array(map( lambda x: mywavelet.dwt(x,tensor=mytensor), data_float ))
interdims = encoded[0].shape
print "interdims:", interdims
compression = find_cutoff_nd( encoded, mycompress )
dicks = map( lambda x: mat_3d_to_dict( x, compression, mythreshfunc ), encoded )
print map( lambda x: compress( x[0], x[1] ), zip( dicks, encoded ) )
lists = map( lambda x: dict_to_mat_3d( x, interdims[0], interdims[1], interdims[2]), dicks )
lists = map( lambda x: np.array(x), lists )
decoded = map( lambda x: mywavelet.idwt(x,tensor=mytensor), lists )
sliced = map( lambda x: np.rint(x).astype( data[0].dtype ), decoded )
print PSNR( np.array(data), np.array(sliced) )
frames, rows, cols = sliced[0].shape
channels = len(sliced)
frame_list = [] #np.zeros((frames, rows, cols, channels))
#maak van lijst van channels, een lijst van frames
for i in range( frames ):
frame = np.zeros((channels, rows, cols))
for j in range( channels ):
frame[j,:,:] = sliced[j][i,:,:]
frame_list.append(frame)
print "converting gifs"
for i, frame in enumerate( frame_list ):
print "converting frame %i" % i
fn = ''.join(["gif/temp.", str(i).zfill(3), ".png"])
mat5img(frame, (rows, cols) ).save( fn )
os.system(' '.join(["convert", fn, fn.replace("png", "gif")]))
if myoutfile == None:
filename = plaat.replace('.','_new.')
else:
filename = myoutfile
print "Saving gif to",filename
os.system(''.join(
["gifsicle -d ",
str(mydelay),
" --loopcount=forever -m ",
"gif/temp.*.gif"
," > ",filename]))
print "removing temp files"
call(["rm"]+list(glob('gif/temp.*.gif')))
call(["rm"]+list(glob('gif/temp.*.png')))
def main(infile):
global mycompress,mywavelet,myoutfile,mytensor,mythreshfunc
data, dim = img3mat(infile)
data = map(lambda x: np.array(x), data)
data_float = map(lambda x: x.astype('float'),data)
print "Starting encoding of image %s" % (argv[1])
print "TensorProduct =",mytensor
print "Threshold func =",mythreshfunc
print "Image has dimensions : %i,%i" % dim
print
encoded = map(lambda x: mywavelet.dwt(x,tensor=mytensor),data_float)
interdims = (len(encoded[0]),len(encoded[0][0]))
print "Encoding done"
# print_dimensions_matlist("Encoded dims:",encoded)
print "Using dimensions:",interdims
print
""" Temporary bypass to check for problems in main algorithm -- Confirmed """
cutoff = find_cutoff_nd( encoded, mycompress )
print "Converted to dictionaries:"
print "Compression cutoff is %f" % cutoff
dicks = map(lambda x: mat5dict(x,cutoff,mythreshfunc),encoded)
print "Dictionary lengths:"
for x in dicks:
print len(x)
print "Compression factor per channel:"
print map(lambda x: compress(x[0],x[1]), zip(dicks,encoded))
print
print "Converting dictionaries to matrices"
print "Dimensions are %i,%i" % interdims
lists = map(lambda x: dict5mat(x,interdims[0],interdims[1]),dicks)
lists = map(lambda x: np.array(x), lists)
print "Starting decoding"
decoded = map(lambda x: mywavelet.idwt(x,tensor=mytensor),lists)
print "Decoding done"
# print_dimensions_matlist("Decoded dimensions are:",decoded)
# print
sliced = map(lambda x: np.rint(x).astype(data[0].dtype),decoded)
# print_dimensions_matlist("Sliced dimensions are:",sliced)
if myoutfile == None:
filename = infile.replace('.','_new.')
else:
filename = myoutfile
print "Saving to",filename
mat5img(sliced,dim).save(filename)
