def load_bow(booksize, files, mute=False):
bow = np.zeros([len(files), booksize])
cnt = -1
for impath in files:
cnt = cnt + 1
if not mute:
print '\r' + str(cnt) + '/' + str(len(files)) + '): ' + impath,
filpat, filnam, filext = tools.fileparts(impath)
filpat2, filnam2, filext2 = tools.fileparts(filpat)
bow[cnt, :] = week56.load_bow('../../data/bow_objects/codebook_' + str(booksize) + '/' + filnam2 + '/' + filnam + '.pkl')
if not mute:
print ''
return bow
def compute_sift(impath, edge_thresh = 10, peak_thresh = 5):
params = '--edge-thresh ' + str(edge_thresh) + ' --peak-thresh ' + str(peak_thresh)
im1 = Image.open(impath).convert('L')
filpat1, filnam1, filext1 = tools.fileparts(impath)
temp_im1 = 'tmp_' + filnam1 + '.pgm'
im1.save(temp_im1)
import struct
is_64bit = struct.calcsize('P') * 8 == 64
if platform.system() == 'Windows' and is_64bit == True:
sift_exec = '..\\..\\external\\vlfeat-0.9.17\\bin\\win64\\sift.exe'
command = sift_exec + ' \'' + os.getcwd() + '\\' + temp_im1 + '\' --output \'' + os.getcwd() + '\\' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Windows' and is_64bit == False:
sift_exec = '..\\..\\external\\vlfeat-0.9.17\\bin\\win32\\sift.exe'
command = sift_exec + ' \'' + os.getcwd() + '\\' + temp_im1 + '\' --output \'' + os.getcwd() + '\\' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Linux':
sift_exec = '..//..//external//vlfeat-0.9.17//bin//glnxa64//sift'
command = sift_exec + ' \'' + os.getcwd() + '//' + temp_im1 + '\' --output \'' + os.getcwd() + '//' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Darwin':
sift_exec = '..//..//external//vlfeat-0.9.17//bin//maci64//sift'
command = sift_exec + ' \'' + os.getcwd() + '//' + temp_im1 + '\' --output \'' + os.getcwd() + '//' + filnam1 + '.sift.output' + '\' ' + params
os.system(command)
frames, sift = read_sift_from_file(filnam1 + '.sift.output')
os.remove(temp_im1)
os.remove(filnam1 + '.sift.output')
return frames, sift
def compute_sift(impath, edge_thresh=10, peak_thresh=5):
global CACHING, WRITTEN_TO_CACHE, RENEW_CACHE
params = ('--edge-thresh '
+ str(edge_thresh)
+ ' --peak-thresh '
+ str(peak_thresh))
im1 = Image.open(impath).convert('L')
if CACHING and not RENEW_CACHE:
if im1 in cache['compute_sift']:
print "retrieving from cache: 'compute_sift'"
return cache['compute_sift'][hash(im1)]
filpat1, filnam1, filext1 = tools.fileparts(impath)
temp_im1 = 'tmp_' + filnam1 + '.pgm'
im1.save(temp_im1)
import struct
is_64bit = struct.calcsize('P') * 8 == 64
if platform.system() == 'Windows' and is_64bit:
sift_exec = '..\\..\\external\\vlfeat-0.9.17\\bin\\win64\\sift.exe'
command = sift_exec + ' \'' + os.getcwd() + '\\' + temp_im1 + '\' --output \'' + os.getcwd() + '\\' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Windows' and not is_64bit:
sift_exec = '..\\..\\external\\vlfeat-0.9.17\\bin\\win32\\sift.exe'
command = sift_exec + ' \'' + os.getcwd() + '\\' + temp_im1 + '\' --output \'' + os.getcwd() + '\\' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Linux':
sift_exec = '..//..//external//vlfeat-0.9.17//bin//glnxa64//sift'
command = sift_exec + ' \'' + os.getcwd() + '//' + temp_im1 + '\' --output \'' + os.getcwd() + '//' + filnam1 + '.sift.output' + '\' ' + params
elif platform.system() == 'Darwin':
sift_exec = '..//..//external//vlfeat-0.9.17//bin//maci64//sift'
command = sift_exec + ' \'' + os.getcwd() + '//' + temp_im1 + '\' --output \'' + os.getcwd() + '//' + filnam1 + '.sift.output' + '\' ' + params
os.system(command)
frames, sift = read_sift_from_file(filnam1 + '.sift.output')
os.remove(temp_im1)
os.remove(filnam1 + '.sift.output')
if CACHING:
cache['compute_sift'][im1] = frames, sift
WRITTEN_TO_CACHE = True
return frames, sift
####
# Get object data
files, labels, label_names, unique_labels, trainset, testset = week56.get_objects_filedata()
####
###
# Load Bag-Of-Words
###
C = 100
bow = np.zeros([len(files), C])
cnt = -1
for impath in files:
cnt = cnt + 1
print str(cnt) + '/' + str(len(files)) + '): ' + impath
filpat, filnam, filext = tools.fileparts(impath)
filpat2, filnam2, filext2 = tools.fileparts(filpat)
bow[cnt, :] = week56.load_bow('../../data/bow_objects/codebook_' + str(C) + '/' + filnam2 + '/' + filnam + '.pkl')
###############################################################################
# Q1: IMPLEMENT HERE kNN CLASSIFIER.
###############################################################################
# Normalize Bag-Of-Words
for i in range(len(files)):
bow[i] = tools.normalizeL1(bow[i])
# k-NN Classifier
dist = np.zeros([len(testset),len(trainset)])
for j in range(len(testset)):
for k in range(len(trainset)):
Results['ItemDuration'] = Results['ItemDuration'] / ts
# Item #
Results['Item'] = pl.arange(1, len(Results['Time']) + 1)
nitems = len(Results['Item'])
Results['Activation'] = pl.zeros(nitems)
Results['Amplitude'] = pl.zeros(nitems)
Results['MeanHR'] = pl.zeros(nitems)
Results['StdHR'] = pl.zeros(nitems)
onset = EDA['Onset'] * SamplingRate
beat = copy(BVP['Onset']) # beat=BVP['ZeroDerivative']/DownSampling
beat /= DownSampling
fparts = tls.fileparts(fname)
rdir = tls.fullfile('../results', fparts[-2])
if not os.path.isdir(rdir): os.mkdir(rdir)
t = pl.arange(len(EDA['Signal'])) / ts / 60.
item = 0
for i in pl.arange(len(idx)):
blbl = 'Block' + str(i + 1)
pl.figure()
y = EDA['Signal'][Results[blbl][0]:Results[blbl][-1]]
ti = t[Results[blbl][0]:Results[blbl][-1]]
