def load_lfw_dataset():
fin=open('data/Foreign_Dataset/LFW_model_5_feature.txt','r')
lines=fin.read().splitlines()
data_num=len(lines)/2
dataset=Dataset()
index=3
gt=dict()
num=0
gt_dict=dict()
for i in xrange(data_num):
dataset.imageNameList.append(['data/Foreign_Dataset/LFW_align/'+lines[i*2]])
dataset.rect.append([0,0,178,218])
name=lines[i*2].split('/')[0]
if gt_dict.has_key(name):
gt_dict[name].append(i)
else:
gt_dict[name]=list()
if gt.has_key(name):
dataset.imgID.append(gt[name])
else:
gt[name]=index
index+=1
dataset.imgID.append(gt[name])
dataset.feature.append(map(float,lines[i*2+1].split()))
dataset.size+=1
#dataset.computeAffinity()
dataset.Quality=[0.99 for i in xrange(dataset.size)]
return dataset,gt_dict
def load_test_data_set(filename):
fin = open(filename, 'r')
lines = fin.read().splitlines()
#data_num = len(lines) / 3
data_num = len(lines) / 4
dataset = Dataset()
dataset.Quality=[]
for i in xrange(data_num):
'''
dataset.imageNameList.append([lines[i * 3]])
dataset.imgID.append(int(lines[i * 3 + 1]))
dataset.feature.append(map(float, lines[i * 3 + 2].split()))
'''
dataset.imageNameList.append([lines[i * 4]])
dataset.rect.append(map(int,lines[i*4+1].split()))
dataset.imgID.append(int(lines[i * 4 + 2]))
dataset.feature.append(map(float, lines[i * 4 + 3].split()))
#dataset.feature[-1]=dataset.feature[-1][:128]
#dataset.Quality.append(1.0)
dataset.size += 1
dataset.computeAffinity()
dataset.computeQuality()
'''
#finetune Quality
quality_path='data/Quality/'+filename.split('/')[-1]
fin_q=open(quality_path,'r')
data_q=fin_q.read().splitlines()
fin_q.close()
dataset.Quality=map(float,data_q)
if len(dataset.Quality)!=dataset.size:
return None
'''
return dataset
def load_LFW_dataset(filepath):
fin=open(filepath,'r')
lines=fin.read().splitlines()
data_num=len(lines)/3
dataset=Dataset()
for i in xrange(data_num):
dataset.imageNameList.append([lines[i*3]])
dataset.imgID.append(int(lines[i*3+1]))
dataset.rect.append([0,0,178,218])
dataset.feature.append(map(float,lines[i*3+2].split()))
dataset.size+=1
dataset.computeAffinity()
#dataset.computeQuality()
dataset.Quality=[0.9 for i in xrange(dataset.size)]
return dataset
def load_nongt_nonquality(filename):
fin = open(filename, 'r')
lines = fin.read().splitlines()
data_num = len(lines) / 3
dataset = Dataset()
for i in xrange(data_num):
dataset.imageNameList.append([lines[i * 3]])
dataset.rect.append(map(int,lines[i*3+1].split()))
dataset.feature.append(map(float, lines[i * 3 + 2].split()))
dataset.imgID.append(0)
dataset.size += 1
dataset.computeAffinity()
dataset.Quality=[0.99 for i in xrange(dataset.size)]
#dataset.computeQuality()
return dataset
def load_HP_dataset(noise=False):
fin=open('/media/heyue/8d1c3fac-68d3-4428-af91-bc478fbdd541/ClusterResearch/clusterQNet/data/HP_model_5_feature.txt','r')
lines=fin.read().splitlines()
data_num=len(lines)/5
dataset=Dataset()
for i in xrange(data_num):
dataset.imageNameList.append(['data/HP/'+lines[i*5]])
dataset.rect.append([0,0,100,100])
dataset.imgID.append(int(lines[i*5+4]))
dataset.feature.append(map(float,lines[i*5+2].split()))
dataset.size+=1
if dataset.imgID[-1]==1 and noise==False:
dataset.imageNameList.pop()
dataset.rect.pop()
dataset.imgID.pop()
dataset.feature.pop()
dataset.size-=1
dataset.computeAffinity()
#dataset.computeQuality()
dataset.Quality=[1.0 for i in xrange(dataset.size)]
return dataset
