def gendata(
data_path,
label_path,
data_out_path,
label_out_path,
num_person_in=1, #observe the first 5 persons
num_person_out=1, #then choose 2 persons with the highest score
max_frame=300):
feeder = Feeder_kinetics(data_path=data_path,
label_path=label_path,
num_person_in=num_person_in,
num_person_out=num_person_out,
window_size=max_frame)
sample_name = feeder.sample_name
sample_label = []
fp = open_memmap(data_out_path,
dtype='float32',
mode='w+',
shape=(len(sample_name), 3, max_frame, 18,
num_person_out))
for i, s in enumerate(sample_name):
data, label = feeder[i]
print_toolbar(
i * 1.0 / len(sample_name),
'({:>5}/{:<5}) Processing data: '.format(i + 1, len(sample_name)))
fp[i, :, 0:data.shape[1], :, :] = data
sample_label.append(label)
with open(label_out_path, 'wb') as f:
pickle.dump((sample_name, list(sample_label)), f)
def gendata(
data_path,
label_path,
data_out_path,
label_out_path,
num_person_in=1, #observe the first 5 persons
num_person_out=1, #then choose 2 persons with the highest score
max_frame=501):
feeder = Feeder_kinetics(data_path=data_path,
label_path=label_path,
num_person_in=num_person_in,
num_person_out=num_person_out,
window_size=max_frame)
print("feeder: ", feeder)
sample_name = feeder.sample_name
print("sample_name: ", sample_name) # 就是每个视频对应的json文件，例如'q8z5Zd5egBQ.json'
sample_label = []
fp = open_memmap(data_out_path,
dtype='float32',
mode='w+',
shape=(len(sample_name), 3, max_frame, 6, num_person_out))
print("type(fp)", type(fp))
for i, s in enumerate(sample_name):
print("i, s:", i, s)
data, label = feeder[i]
print("data, label:", type(data), type(label))
print("data.shape:", data.shape)
print_toolbar(
i * 1.0 / len(sample_name),
'({:>5}/{:<5}) Processing data: '.format(i + 1, len(sample_name)))
fp[i, :, 0:data.shape[1], :, :] = data
sample_label.append(label)
with open(label_out_path, 'wb') as f:
pickle.dump((sample_name, list(sample_label)), f)
def gendata(
data_path,
label_path,
data_out_path,
label_out_path,
num_person_in=5, #observe the first 5 persons
num_person_out=2, #then choose 2 persons with the highest score
max_frame=300):
feeder = Feeder_kinetics(
data_path=data_path,
label_path=label_path,
num_person_in=num_person_in,
num_person_out=num_person_out,
window_size=max_frame)
sample_name = feeder.sample_name
sample_label = []
'''fp = open_memmap(
data_out_path,
dtype='float32',
mode='w+',
shape=(len(sample_name), 3, max_frame, 18, num_person_out))'''
skeleton_id_1_origin = '408092820'
skeleton_id_2_origin = '975158737'
for i, s in enumerate(sample_name):
data, label = feeder[i]
print_toolbar(i * 1.0 / len(sample_name),
'({:>5}/{:<5}) Processing data: '.format(
i + 1, len(sample_name)))
#fp[i, :, 0:data.shape[1], :, :] = data
if label < 10:
action_id = '00' + str(label)
elif label >=10 and label < 100:
action_id = '0' + str(label)
else:
action_id = str(label)
skeleton_id_1 = skeleton_id_1_origin + str(i)
skeleton_id_2 = skeleton_id_2_origin + str(i)
file_name = data_out_path + '/' + action_id + '_' + s.split('.')[0] + '.skeleton'
f = open(file_name, "w+")
f.write(str(300) + '\n')
for j in range(300):
'''f.write(str(2) + '\n')
f.write(skeleton_id_1 + '\n')
f.write(str(18) + '\n')
for k in range(18):
output_data_1 = str(data[0][j][k][0]) + " " + str(data[1][j][k][0]) + " " + str(data[2][j][k][0]) + ' 0 0 0 0 0 0 0 0'
f.write(str(output_data_1) + '\n')
f.write(skeleton_id_2 + '\n')
f.write(str(18) + '\n')
for l in range(18):
output_data_2 = str(data[0][j][l][1]) + " " + str(data[1][j][l][1]) + " " + str(data[2][j][l][1]) + ' 0 0 0 0 0 0 0 0'
f.write(str(output_data_2) + '\n')'''
f.write(str(2) + '\n')
f.write(skeleton_id_1 + '\n')
f.write(str(25) + '\n')
for k in range(2):
body_1_joint_1 = str((data[0][j][8][k] + data[0][j][11][k])/2) + " " + str((data[1][j][8][k] + data[1][j][11][k])/2) + " " + str((data[2][j][8][k] + data[2][j][11][k])/2) + ' 0 0 0 0 0 0 0 0'
body_1_joint_2 = str(((data[0][j][8][k] + data[0][j][11][k])/2 + data[0][j][1][k])/2) + " " + str(((data[1][j][8][k] + data[1][j][11][k])/2 + data[1][j][1][k])/2) + " " + str(((data[2][j][8][k] + data[2][j][11][k])/2 + data[2][j][1][k])/2) + ' 0 0 0 0 0 0 0 0'
body_1_joint_3 = str((data[0][j][0][k] + data[0][j][1][k])/2) + " " + str((data[1][j][0][k] + data[1][j][1][k])/2) + " " + str((data[2][j][0][k] + data[2][j][1][k])/2) + ' 0 0 0 0 0 0 0 0'
body_1_joint_4 = str(data[0][j][0][k]) + " " + str(data[1][j][0][k]) + " " + str(data[2][j][0][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_5 = str(data[0][j][5][k]) + " " + str(data[1][j][5][k]) + " " + str(data[2][j][5][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_6 = str(data[0][j][6][k]) + " " + str(data[1][j][6][k]) + " " + str(data[2][j][6][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_7 = str(data[0][j][7][k]) + " " + str(data[1][j][7][k]) + " " + str(data[2][j][7][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_8 = str(data[0][j][7][k]) + " " + str(data[1][j][7][k]) + " " + str(data[2][j][7][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_9 = str(data[0][j][2][k]) + " " + str(data[1][j][2][k]) + " " + str(data[2][j][2][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_10 = str(data[0][j][3][k]) + " " + str(data[1][j][3][k]) + " " + str(data[2][j][3][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_11 = str(data[0][j][4][k]) + " " + str(data[1][j][4][k]) + " " + str(data[2][j][4][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_12 = str(data[0][j][4][k]) + " " + str(data[1][j][4][k]) + " " + str(data[2][j][4][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_13 = str(data[0][j][11][k]) + " " + str(data[1][j][11][k]) + " " + str(data[2][j][11][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_14 = str(data[0][j][12][k]) + " " + str(data[1][j][12][k]) + " " + str(data[2][j][12][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_15 = str(data[0][j][13][k]) + " " + str(data[1][j][13][k]) + " " + str(data[2][j][13][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_16 = str(data[0][j][13][k]) + " " + str(data[1][j][13][k]) + " " + str(data[2][j][13][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_17 = str(data[0][j][8][k]) + " " + str(data[1][j][8][k]) + " " + str(data[2][j][8][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_18 = str(data[0][j][9][k]) + " " + str(data[1][j][9][k]) + " " + str(data[2][j][9][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_19 = str(data[0][j][10][k]) + " " + str(data[1][j][10][k]) + " " + str(data[2][j][10][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_20 = str(data[0][j][10][k]) + " " + str(data[1][j][10][k]) + " " + str(data[2][j][10][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_21 = str(data[0][j][1][k]) + " " + str(data[1][j][1][k]) + " " + str(data[2][j][1][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_22 = str(data[0][j][7][k]) + " " + str(data[1][j][7][k]) + " " + str(data[2][j][7][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_23 = str(data[0][j][7][k]) + " " + str(data[1][j][7][k]) + " " + str(data[2][j][7][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_24 = str(data[0][j][4][k]) + " " + str(data[1][j][4][k]) + " " + str(data[2][j][4][k]) + ' 0 0 0 0 0 0 0 0'
body_1_joint_25 = str(data[0][j][4][k]) + " " + str(data[1][j][4][k]) + " " + str(data[2][j][4][k]) + ' 0 0 0 0 0 0 0 0'
f.write(body_1_joint_1 + '\n')
f.write(body_1_joint_2 + '\n')
f.write(body_1_joint_3 + '\n')
f.write(body_1_joint_4 + '\n')
f.write(body_1_joint_5 + '\n')
f.write(body_1_joint_6 + '\n')
f.write(body_1_joint_7 + '\n')
f.write(body_1_joint_8 + '\n')
f.write(body_1_joint_9 + '\n')
f.write(body_1_joint_10 + '\n')
f.write(body_1_joint_11 + '\n')
f.write(body_1_joint_12 + '\n')
f.write(body_1_joint_13 + '\n')
f.write(body_1_joint_14 + '\n')
f.write(body_1_joint_15 + '\n')
f.write(body_1_joint_16 + '\n')
f.write(body_1_joint_17 + '\n')
f.write(body_1_joint_18 + '\n')
f.write(body_1_joint_19 + '\n')
f.write(body_1_joint_20 + '\n')
f.write(body_1_joint_21 + '\n')
f.write(body_1_joint_22 + '\n')
f.write(body_1_joint_23 + '\n')
f.write(body_1_joint_24 + '\n')
f.write(body_1_joint_25 + '\n')
if k != 1:
f.write(skeleton_id_2 + '\n')
f.write(str(25) + '\n')
f.close()
#sample_label.append(label)
'''with open(label_out_path, 'wb') as f: