def get_label_data(label):
files = os.listdir(label)
test_mfcc = [
MFCC.get_mfcc(os.path.join(label, f)) for f in files
if f.endswith("wav")
]
return test_mfcc
def get_label_data(directory):
files = os.listdir(directory)
# Randomized Test and Train set
test_files = rand.sample(files, TEST_SIZE)
train_mfcc = []
test_mfcc = []
for f in test_files:
if f.endswith("wav"):
test_mfcc.append(MFCC.get_mfcc(os.path.join(directory, f)))
files.pop(files.index(f))
train_mfcc = [MFCC.get_mfcc(os.path.join(directory, f)) for f in files if f.endswith("wav")]
return train_mfcc, test_mfcc
def runHMM(file_path):
models = {}
for label in CLASS_LABELS:
with open(os.path.join("Models", label + ".pkl"), "rb") as file:
models[label] = pk.load(file)
with open("Models/kmeans.pkl", "rb") as file:
kmeans = pk.load(file)
sound_mfcc = MFCC.get_mfcc(file_path)
sound_mfcc = kmeans.predict(sound_mfcc).reshape(-1, 1)
evals = {
cname: model.score(sound_mfcc, [len(sound_mfcc)])
for cname, model in models.items()
}
conclusion = max(evals.keys(), key=(lambda k: evals[k]))
return evals, conclusion
