def match_cosine_similarity(self, image_path, topn=5):
# Compare image_path with other images
features = fe.extract_features(image_path)
img_distances = self.cosine_similarity(np.array(features))
# Slice list with len() = topn after sorted descending
nearest_img_index = np.argsort(img_distances)[::-1][:topn].tolist()
nearest_img_paths = self.names[nearest_img_index].tolist()
return nearest_img_paths, img_distances[nearest_img_index].tolist()
def getDataset(self,df,datasetType):
features = []
tag = []
for index,row in df.iterrows():
if datasetType is 'test':
# data.append(features.extract_features(row['Sentence1'],row['Sentence2'],row['GoldTag']))
features.append(featureExtraction.extract_features(
row['Sentence1'], row['Sentence2'])
)
else:
features.append(
featureExtraction.extract_features(
row['Sentence1'], row['Sentence2']
))
tag.append(row['GoldTag'])
if datasetType is 'train' or 'dev':
label = pd.DataFrame(tag, columns = ['gold_tag'])
dataset = pd.DataFrame(features, columns = ['Jaccard','bow','cosine','hyper_ratio',
'hypo_ratio','holo_ratio','mero_ratio'])
return dataset, label
def train():
display = tk.Text(master=window, height=8, width=40)
display.config(font=("Helvetica"))
display.grid(columnspan=2, row=5, sticky='e')
warnings.filterwarnings("ignore")
# path to training data
source = "data/"
# path where training speakers will be saved
dest = "models/"
train_file = "data.txt"
file_paths = open(train_file, 'r')
count = 1
# Extracting features for each speaker (5 files per speakers)
features = np.asarray(())
for path in file_paths:
path = path.strip()
# print path
# display.insert(tk.END, path)
# display.insert(tk.END, "\n")
# read the audio
sr, audio = read(source + path)
# extract 40 dimensional MFCC & delta MFCC features
vector = extract_features(audio, sr)
if features.size == 0:
features = vector
else:
features = np.vstack((features, vector))
# when features of 5 files of speaker are concatenated, then do model training
if count == 5:
gmm = GMM(n_components=16,
n_iter=200,
covariance_type='diag',
n_init=3)
gmm.fit(features)
# dumping the trained gaussian model
picklefile = path.split("/")[0] + ".gmm"
# print picklefile
# f = open(dest + picklefile, 'w+')
cPickle.dump(gmm, open(dest + picklefile, 'w+'))
# print '+ modeling completed for word:', picklefile, " with data point = ", features.shape
phrase = "Modeling completed: " + picklefile
display.insert(tk.END, phrase)
display.insert(tk.END, "\n")
features = np.asarray(())
count = 0
count = count + 1
import sys
from sklearn import svm
import featureExtraction
from readData import CorpusReader
from sklearn import svm
# class model:
if __name__ == "__main__":
# trainSet = sys.argv[1]
# datasetType = sys.argv[2]
s1 = "The young lady enjoys listening to the guitar."
s2 = "The young lady enjoys playing the guitar."
tag = 4
# extract features
features = featureExtraction.extract_features(s1, s2)
dataset = pd.Dataframe(features,
columns=[
'Jaccard', 'bow', 'cosine', 'hyper_ratio',
'hypo_ratio', 'holo_ratio', 'mero_ratio'
])
label = pd.Dataframe(tag, column=['gold_tag'])
# call model and save model
# evaluate
def test():
display = tk.Text(master=window, height=8, width=40)
display.config(font=("Helvetica"))
display.grid(columnspan=2, row=5, sticky='e')
WAVE_OUTPUT_FILENAME = "test/test"
# #path to training data
source = "data/"
modelpath = "models/"
# test_file = "development_set_test.txt"
# file_paths = open(test_file,'r')
#
gmm_files = [
os.path.join(modelpath, fname) for fname in os.listdir(modelpath)
if fname.endswith('.gmm')
]
# Load the Gaussian gender Models
models = [cPickle.load(open(fname, 'r')) for fname in gmm_files]
speakers = [fname.split("/")[-1].split(".gmm")[0] for fname in gmm_files]
# Read the test directory and get the list of test audio files
p = pyaudio.PyAudio()
stream = p.open(format=FORMAT,
channels=CHANNELS,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
# print("* recording")
frames = []
for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
data = stream.read(CHUNK)
frames.append(data)
# print("* done recording")
stream.stop_stream()
stream.close()
p.terminate()
wf = wave.open(WAVE_OUTPUT_FILENAME + ".wav", 'wb')
wf.setnchannels(CHANNELS)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(b''.join(frames))
wf.close()
sr, audio = read("test/test.wav")
vector = extract_features(audio, sr)
log_likelihood = np.zeros(len(models))
for i in range(len(models)):
gmm = models[i] # checking with each model one by one
scores = np.array(gmm.score(vector))
log_likelihood[i] = scores.sum()
winner = np.argmax(log_likelihood)
# print "\tDETECTED AS: ", speakers[winner]
display.insert(tk.END, "DETECTED AS: " + speakers[winner])
execute(speakers[winner])