def __init__(self, avi_file=None, wav_file=None):
if wav_file:
self.wav_file = wav_file
elif avi_file:
self.avi_file = avi_file
self.wav_file = avi_file.split('.')[0] + '.wav'
subprocess.call([
'ffmpeg', '-y', '-i',
str(self.avi_file), '-ab', '160k', '-ac', '2', '-ar', '44100',
'-vn',
str(self.wav_file)
])
else:
raise Exception()
self.model = model
self.subjectWav = SubjectWav(str(self.wav_file))
def __init__(self, avi_file=None, wav_file=None):
if wav_file:
self.wav_file = wav_file
elif avi_file:
self.avi_file = avi_file
self.wav_file = avi_file.split('.')[0] + '.wav'
subprocess.call(['ffmpeg', '-y', '-i', str(self.avi_file), '-ab', '160k', '-ac', '2', '-ar', '44100', '-vn', str(self.wav_file)])
else:
raise Exception()
self.model = model
self.subjectWav = SubjectWav(str(self.wav_file))
class HeartRateGenerator(object):
def __init__(self, avi_file=None, wav_file=None):
if wav_file:
self.wav_file = wav_file
elif avi_file:
self.avi_file = avi_file
self.wav_file = avi_file.split('.')[0] + '.wav'
subprocess.call([
'ffmpeg', '-y', '-i',
str(self.avi_file), '-ab', '160k', '-ac', '2', '-ar', '44100',
'-vn',
str(self.wav_file)
])
else:
raise Exception()
self.model = model
self.subjectWav = SubjectWav(str(self.wav_file))
# Yield and average heartrate over 4 seconds for every second in a audio
# file except the first 4
def gen_heartrates(self):
window = []
time = 0
while True:
try:
spectrogram = self.subjectWav.get_spectrogram(time)
spectrogram = spectrogram[2]
spectrogram = np.array([[spectrogram]])
spectrogram = normalizer.normalize_data(spectrogram)
spectrogram = sliceToTimeSeries(np.array(spectrogram))
heartrate = self.model.predict(spectrogram)[0][0]
heartrate = normalizer.unnormalize_bpm(heartrate)
heartrate = heartrate if str(heartrate) != 'nan' else None
window.append(heartrate)
if len(window) is 5:
window = window[1:]
if len(filter(lambda x: x, window)) is 4:
yield np.mean(window)
else:
yield None
except Exception:
break
time += 1
class HeartRateGenerator(object):
def __init__(self, avi_file=None, wav_file=None):
if wav_file:
self.wav_file = wav_file
elif avi_file:
self.avi_file = avi_file
self.wav_file = avi_file.split('.')[0] + '.wav'
subprocess.call(['ffmpeg', '-y', '-i', str(self.avi_file), '-ab', '160k', '-ac', '2', '-ar', '44100', '-vn', str(self.wav_file)])
else:
raise Exception()
self.model = model
self.subjectWav = SubjectWav(str(self.wav_file))
# Yield and average heartrate over 4 seconds for every second in a audio
# file except the first 4
def gen_heartrates(self):
window = []
time = 0
while True:
try:
spectrogram = self.subjectWav.get_spectrogram(time)
spectrogram = spectrogram[2]
spectrogram = np.array([[spectrogram]])
spectrogram = normalizer.normalize_data(spectrogram)
spectrogram = sliceToTimeSeries(np.array(spectrogram))
heartrate = self.model.predict(spectrogram)[0][0]
heartrate = normalizer.unnormalize_bpm(heartrate)
heartrate = heartrate if str(heartrate) != 'nan' else None
window.append(heartrate)
if len(window) is 5:
window = window[1:]
if len(filter(lambda x: x, window)) is 4:
yield np.mean(window)
else:
yield None
except Exception:
break
time += 1