def main(csv_file):
# Check input file whether being csv.
if csv_file.split(".")[-1] != "csv":
print("Input file is not a csv file.")
return
#run_translator("data/es-en.csv")
#calculate_ribes("result/es-en.csv")
calculate_features("result/es-en.csv")
print("Success on executing main!!")
def get_features(data, params=Constants()):
excluded = 0
overall = 0
IPy = False
if "IPython.display" in sys.modules.keys():
from IPython.display import clear_output
IPy = True
for di, d in enumerate(data):
if di % 100 == 0:
if IPy:
clear_output()
print(di, ' out of ', len(data))
st_features = calc_feat.calculate_features(d['signal'],
params.framerate, None).T
x = []
y = []
for f in st_features:
overall += 1
if f[1] > 1.e-4:
x.append(f)
y.append(d['emotion'])
else:
excluded += 1
x = np.array(x, dtype=float)
y = np.array(y)
save_sample(x, y, params.path_to_features + d['id'] + '.csv')
return overall, excluded
def emotion_predict(path):
wav = open_wav(path)
(nchannels, sampwidth, framerate, nframes, comptype,
compname), samples = wav
duration = nframes / framerate
print('channel count', nchannels)
left = samples[0::nchannels]
print('left', left)
print('left shape', left.shape)
# right = samples[1::nchannels]
# left_audio = left[int(start * framerate):int(end * framerate)]
st_features = cf.calculate_features(left, framerate, None).T
x = []
for f in st_features:
if f[1] > 1.e-4:
x.append(f)
x = np.array(x, dtype=float)
tx = []
tx.append(np.array(x, dtype=float))
tx = np.array(tx, dtype=float)
tx = pad_sequence(tx, 32)
model = load_model('iemocap_acc0.561715.h5', compile=False)
available_emotions = ['生气', '激动', '紧张', '悲伤']
predictResult = model.predict(tx)
#emotion = available_emotions[np.argmax(predictResult)]
#prob = predictResult[0][np.argmax(predictResult)]
return predictResult[0].tolist()
def driver_trips(driver_number):
number_of_trips = 200
df = np.zeros((200, 18), dtype=np.float32)
for i in range(1, number_of_trips + 1, 1):
filename = 'drivers/' + str(driver_number) + '/' + str(i) + '.csv'
fdict = calculate_features.calculate_features(filename)
df[i - 1] = fdict
#for col in range(df.shape[1]):
#df[:,col] = (df[:,col] - df[:,col].mean())/(df[:,col].max() - df[:,col].min())
return df
def train_model(images, mask_list, k_size, probability):
"""
Train model with list of images
"""
logging.info('Calculating, normalizing feature vectors for %d image(s)',
len(images))
vectors_list = [
calculate_features(x.image, x.fov_mask, mask_list, k_size)
for x in images
]
truth_list = [x.truth for x in images]
logging.info('Training model with %d image(s)', len(images))
svm.train(vectors_list, truth_list,
probability) # Train SVM, lengthy process
def get_features(data, save=True, path='samples/l_', mode='calculate'):
if mode == 'calculate':
for di, d in enumerate(data):
print di, ' out of ', len(data)
st_features = cf.calculate_features(d['left'], None).T
x = []
y = []
for f in st_features:
if f[1] > 1.e-4:
x.append(f)
y.append(d['emotion'])
x = np.array(x, dtype=float)
y = np.array(y)
if save:
save_sample(x, y, path + d['id'] + '.csv')
return x, y
def get_features(data, save=True, path='samples/l_', mode='calculate'):
if mode == 'calculate':
for di, d in enumerate(data):
print di, ' out of ', len(data)
st_features = cf.calculate_features(d['left'], None).T
x = []
y = []
for f in st_features:
if f[1] > 1.e-4:
x.append(f)
y.append(d['emotion'])
x = np.array(x, dtype=float)
y = np.array(y)
if save:
save_sample(x, y, path + d['id'] + '.csv')
return x, y
def get_features(data, save=True, path=path_to_samples):
failed_samples = []
for di, d in enumerate(data):
if di%1000 == 0:
print di, ' out of ', len(data)
st_features = cf.calculate_features(d['signal'], framerate, None).T
x = []
y = []
for f in st_features:
if f[1] > 1.e-4:
x.append(f)
y.append(d['emotion'])
x = np.array(x, dtype=float)
y = np.array(y)
if save:
save_sample(x, y, path + d['id'] + '.csv')
return x, y
def get_features(data, save=True, path=path_to_samples):
failed_samples = []
for di, d in enumerate(data):
if di % 1000 == 0:
print di, ' out of ', len(data)
st_features = cf.calculate_features(d['signal'], framerate, None).T
x = []
y = []
for f in st_features:
if f[1] > 1.e-4:
x.append(f)
y.append(d['emotion'])
x = np.array(x, dtype=float)
y = np.array(y)
if save:
save_sample(x, y, path + d['id'] + '.csv')
return x, y
def classify_image(images, mask_list, k_size, save, display):
"""
Classify pixels of a single image
"""
if len(images) > 1:
raise ValueError('Only one image can be classified at once')
logging.info('Calculating, normalizing feature vectors for image')
image = images[0] # First and only member
vectors = calculate_features(image.image, image.fov_mask, mask_list,
k_size)
logging.info('Classifying image pixels')
probabilities, prediction = svm.classify(vectors)
svm.assess(image.truth, prediction)
svm.plot_roc(image.truth, probabilities)
if save:
image_utils.save_image(prediction, 'prediction.png')
logging.info('Saved classified image')
if display:
image_utils.display_image(prediction)
logging.info('Displaying classified image')