def read_class(standardize=False):
"""read featuers into X and labels for instruments family into y. if standardize set to true, then it will return a scaler that is
used to transform the test dataset.
"""
X, y = read_features()
instruments = [ins[0] for ins in y]
X = np.asarray(X)
scaler = None
if standardize:
scaler = preprocessing.StandardScaler().fit(X)
X = scaler.transform(X)
y = np.asarray(instruments)
return X, y, scaler
def read_class(standardize=False):
"""read featuers into X and labels for instruments family into y. if standardize set to true, then it will return a scaler that is
used to transform the test dataset.
"""
X, y = read_features()
instruments = [ins[0] for ins in y]
X = np.asarray(X)
scaler = None
if standardize:
scaler = preprocessing.StandardScaler().fit(X)
X = scaler.transform(X)
y = np.asarray(instruments)
return X, y, scaler
layers=[
Layer("Tanh", units=12),
Layer("Softmax")],
learning_rate=0.005,
n_iter=25)
# gs = GridSearchCV(nn, param_grid={
# 'learning_rate': [0.05, 0.01, 0.005, 0.001],
# 'hidden0__units': [4, 8, 12,100],
# 'hidden0__type': ["Rectifier", "Sigmoid", "Tanh"]})
# gs.fit(X_train, y_train)
# print(gs.best_estimator_)
nn.fit(X_train, y_train)
predicted = nn.predict(X_test).flatten()
labels = y_test
return predicted, labels
if __name__ == "__main__":
features = ['zcr', 'rms', 'sc', 'sr', 'sf','mfcc']
X, y = read_features(features)
X = clean_data(X)
# train_score, test_score = train_nolearn_model(X, y)
# print("train score: ", train_score)
# print("test score: ", test_score)
predicted, true_value = train_sknn(X,y)
print("predicted: ", predicted)
print("true_value: ", true_value)
print("accuracy: ", np.sum(predicted == true_value) / float(len(predicted)))
#the test accuracy is only 38%, need more tunning!!!
def run_cnn(dimension, _load=False):
# intialize the log file for current run of the code
utils.initialize_log()
# read audio files and parse them or simply load from pre-extracted feature files
if _load:
dictionary, tr_mnn_features, tr_mnn_labels, ts_mnn_features, ts_mnn_name_list, tr_cnn_features, tr_cnn_labels, ts_cnn_features, ts_cnn_name_list = read_audio_files(
)
else:
dictionary, tr_mnn_features, tr_mnn_labels, ts_mnn_features, ts_mnn_name_list, tr_cnn_features, tr_cnn_labels, ts_cnn_features, ts_cnn_name_list = features.read_features(
)
if dimension == 1:
# call the 2d convolutional network code here
cnn_1d_probs = train.keras_convolution_1D(tr_cnn_features,
tr_cnn_labels,
ts_cnn_features,
len(dictionary),
training_epochs=50)
# get top three predictions
top3 = cnn_1d_probs.argsort()[:, -3:][:, ::-1]
else:
# call the 2d convolutional network code here
cnn_2d_probs = train.keras_convolution_2D(tr_cnn_features,
tr_cnn_labels,
ts_cnn_features,
len(dictionary),
training_epochs=50)
# get top three predictions
top3 = cnn_2d_probs.argsort()[:, -3:][:, ::-1]
# print the predicted results to a csv file.
utils.print_csv_file(top3, ts_cnn_name_list, dictionary, OUTPUT_CSV)
from sklearn.ensemble import RandomForestClassifier
import random
from features import read_features
def get_class(num):
if num < 0.17:
return 0
elif num < 0.5:
return 1
elif num < 0.83:
return 2
else:
return 3
data, labels = read_features()
order = random.sample(range(len(data)), len(data))
train_idx = order[:14570]
test_idx = order[14570:]
# 2 classes
labels = [round(label) for label in labels]
# 4 classes
labels = [get_class(label) for label in labels]
train_data = []
train_labels = []
test_data = []
test_labels = []
for idx,d in enumerate(data):
if idx in train_idx:
train_data.append(d)
train_labels.append(labels[idx])
print("For C: %f, train_score=%f, test_score=%f" %
(c, train_score, test_score))
print()
def rf_tuning(X, y):
n_trees = [100, 500, 800]
for n in n_trees:
rf = RandomForestClassifier(n_estimators=n)
test_score, train_score, cms = train_model(X, y, rf)
print(rf.feature_importances_)
print("For n_tree: %f, train_score=%f, test_score=%f" %
(n, train_score, test_score))
print(np.sum(cms, axis=0))
print()
if __name__ == "__main__":
features = ['zcr', 'rms', 'sc', 'sr', 'sf', 'mfcc']
X, y = read_features(features)
# print(X.shape)
X = clean_data(X)
#find_best_k(X,y)
rf_tuning(X, y)
#plot learning curve for 8nn
# clf = KNeighborsClassifier(n_neighbors=8, weights = 'distance')
# crossvalidation = cross_validation.StratifiedKFold(y, n_folds=5)
# plot_learning_curve(clf, "knn_learning_curve", X, y, cv= crossvalidation, n_jobs=4)
def main(_load = False):
# intialize the log file for current run of the code
utils.initialize_log()
# read audio files and parse them or simply load from pre-extracted feature files
if _load:
dictionary, tr_mnn_features, tr_mnn_labels, ts_mnn_features, ts_mnn_name_list, tr_cnn_features, tr_cnn_labels, ts_cnn_features, ts_cnn_name_list = read_audio_files()
else:
dictionary, tr_mnn_features, tr_mnn_labels, ts_mnn_features, ts_mnn_name_list, tr_cnn_features, tr_cnn_labels, ts_cnn_features, ts_cnn_name_list = features.read_features()
# print a log message for status update
utils.write_log_msg("starting multi-layer neural network training...")
# use the above extracted features for the training of the model
predictions_top3 = train.train(tr_mnn_features, tr_mnn_labels, ts_mnn_features, tr_cnn_features, tr_cnn_labels, ts_cnn_features, n_classes=len(dictionary))
# print a log message for status update
utils.write_log_msg("outputing prediction results to a csv file...")
# print the predicted results to a csv file.
utils.print_csv_file(predictions_top3, ts_mnn_name_list, dictionary, OUTPUT_CSV)
# print a log message for status update
utils.write_log_msg("done...")
if __name__ == '__main__':
from sys import argv
# argv check / usage info
if len(argv) < 2:
print("Usage: {} image_path".format(argv[0]))
exit(1)
# Read the image
img = cv2.imread(argv[1])
if img is None:
print("Error loading image")
exit(1)
# Get the features from the file
fsize, features = read_features('numbers.feat')
# Get all contours from the image
contours = get_all_contours(img)
# Loop through all contours
for contour in contours:
# Get the symbol and attempt to match it to a known feature
symbol = scale_symbol(img, contour, (fsize,fsize))
ret = recognize_symbol(features, symbol, fsize)
# Output the results
if ret is not None:
print("identified symbol: '{}'".format(ret))
pyplot.imshow(symbol)
pyplot.show()