def main():
images_transformed_path = get_config('images_transformed_path')
with open(images_transformed_path, 'w') as output_file:
writer = csv.writer(output_file, delimiter=',')
training_images_labels_path = get_config('training_images_labels_path')
with open(training_images_labels_path, 'r') as file:
lines = file.readlines()
for line in lines:
print("\n\n" + line.strip())
image_path, image_label = line.split()
# Read the input image.
frame = cv2.imread(image_path)
# `frame` is a HxW numpy ndarray of triplets (pixels), where H and W are
# the dimensions of the input image.
# cv2.imshow("Original", frame)
try:
frame = apply_image_transformation(frame)
write_frame_to_file(frame, image_label, writer)
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation on image path '{}' with the following exception trace:\n{}"
.format(image_path, exception_traceback))
continue
# cv2.waitKey(1000)
cv2.destroyAllWindows()
print "The program completed successfully !!"
def main():
model_name = sys.argv[1]
if model_name not in ['svm', 'logistic', 'knn']:
logger.error("Invalid model-name '{}'!".format(model_name))
return
logger.info("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
'model_{}_serialized_path'.format(model_name))
logger.info(
"Model deserialized from path '{}'".format(model_serialized_path))
for root, dir, filex in os.walk("C:/xampp/htdocs/file_upload_api/files/",
topdown=True):
image_path = "C:/xampp/htdocs/file_upload_api/files/" + filex[0]
break
frame = cv2.imread(image_path)
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
classifier_model = joblib.load(model_serialized_path)
predicted_labels = classifier_model.predict(
frame_flattened.reshape(1, -1))
predicted_label = predicted_labels[0]
print("Predicted labelx={}".format(predicted_label))
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation on image path '{}' with the following exception trace:\n{}"
.format(image_path, exception_traceback))
os.remove(image_path)
cv2.destroyAllWindows()
logger.info("The program completed successfully !!")
def main():
model_name = sys.argv[1]
if model_name not in ['svm', 'logistic', 'knn']:
print("Invalid model-name '{}'!".format(model_name))
return
print("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
"model_{}_serialized_path".format(model_name))
print("Model deserialized from path '{}'".format(model_serialized_path))
camera = cv2.VideoCapture(0)
while True:
ret, frame = camera.read()
cv2.imshow("Recording", frame)
key = cv2.waitKey(1)
if key == ord('q'):
break
elif key == ord('c'):
if not ret:
print("Failed to capture image!")
continue
frame = resize_image(frame, 400)
r = cv2.selectROI(frame)
imCrop = frame[int(r[1]):int(r[1] + r[3]),
int(r[0]):int(r[0] + r[2])]
cv2.imshow("Image", imCrop)
#cv2.imshow("Webcam recording", frame)
frame = imCrop
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
classifier_model = joblib.load(model_serialized_path)
predicted_labels = classifier_model.predict(frame_flattened)
predicted_label = predicted_labels[0]
print("Predicted label = {}".format(predicted_label))
predicted_image = get_image_from_label(predicted_label)
predicted_image = resize_image(predicted_image, 200)
cv2.imshow("Prediction = '{}'".format(predicted_label),
predicted_image)
engine = pyttsx.init()
engine.say("The predicted text is " + str(predicted_label))
engine.runAndWait()
engine.stop()
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation with the following exception trace:\n{}"
.format(exception_traceback))
cv2.destroyAllWindows()
print "The program completed successfully !!"
def mainimage():
font = cv2.FONT_HERSHEY_SIMPLEX
bottomLeftCornerOfText = (10, 50)
fontScale = 1
fontColor = (255, 255, 255)
lineType = 2
model_name = "svm"
if model_name not in ['svm']:
print("Invalid model-name '{}'!".format(model_name))
return
print("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
"model_{}_serialized_path".format(model_name))
print("Model deserialized from path '{}'".format(model_serialized_path))
classifier_model = joblib.load(model_serialized_path)
import tkFileDialog
tk.Tk().withdraw()
path = tkFileDialog.askopenfile()
img = cv2.imread(path.name)
frame = resize_image(img, 400)
image = frame
frame = cv2.resize(img, (50, 50))
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
print classifier_model
predicted_labels = classifier_model.predict(frame_flattened)
print("Predicted label = {}".format(predicted_labels))
cv2.putText(image, str(predicted_labels), bottomLeftCornerOfText, font,
fontScale, fontColor, lineType)
cv2.imshow("frame", image)
key = cv2.waitKey(2)
if key == 27:
cv2.destroyAllWindows()
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation with the following exception trace:\n{}"
.format(exception_traceback))
cv2.waitKey(0)
print "The program completed successfully !!"
def main():
model_name = "svm"
if model_name not in ['svm', 'logistic', 'knn']:
print("Invalid model-name '{}'!".format(model_name))
return
print("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
"model_{}_serialized_path".format(model_name))
print("Model deserialized from path '{}'".format(model_serialized_path))
camera = cv2.VideoCapture(0)
while True:
ret, frame = camera.read()
if not ret:
print("Failed to capture image!")
continue
frame = resize_image(frame, 400)
cv2.imshow("Webcam recording", frame)
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
#print(model_serialized_path)
#model_serialized_path = "C:\\Users\\AanikaRahman\\Documents\\GitHub\\ASL1\\Sign-Language-Recognition\\data\\generated\\output\\svm\\model-serialized-svm.pkl"
classifier_model = joblib.load(model_serialized_path)
#classifier_model = pickle.load(open(model_serialized_path),encoding='latin1')
predicted_labels = classifier_model.predict(frame_flattened)
predicted_label = predicted_labels[0]
print("Predicted label = {}".format(predicted_label))
predicted_image = get_image_from_label(predicted_label)
predicted_image = resize_image(predicted_image, 200)
cv2.imshow("Prediction = '{}'".format(predicted_label),
predicted_image)
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation with the following exception trace:\n{}"
.format(exception_traceback))
cv2.waitKey(2000)
cv2.destroyAllWindows()
cv2.destroyAllWindows()
print("The program completed successfully !!")
def main():
model_name = sys.argv[1]
if model_name not in ['svm', 'logistic', 'knn']:
logger.error("Invalid model-name '{}'!".format(model_name))
return
logger.info("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
'model_{}_serialized_path'.format(model_name))
logger.info(
"Model deserialized from path '{}'".format(model_serialized_path))
testing_images_labels_path = get_config('testing_images_labels_path')
with open(testing_images_labels_path, 'r') as file:
lines = file.readlines()
for line in lines:
if not line:
continue
image_path, image_label = line.split()
frame = cv2.imread(image_path)
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
classifier_model = joblib.load(model_serialized_path)
predicted_labels = classifier_model.predict(
frame_flattened.reshape(1, -1))
predicted_label = predicted_labels[0]
logger.info('"{}" {} ---> {}'.format(image_path, image_label,
predicted_label))
if image_label != predicted_label:
log_msg = "Incorrect prediction '{}' instead of '{}'\n)"
logger.error(log_msg.format(predicted_label, image_label))
cv2.waitKey(5000)
except Exception:
exception_traceback = traceback.format_exc()
logger.error("Error applying image transformation to image "
"'{}'".format(image_path))
logger.debug(exception_traceback)
continue
cv2.destroyAllWindows()
logger.info("The program completed successfully !!")
def main():
model_name = sys.argv[1]
if model_name not in ['svm', 'logistic', 'knn']:
print("Invalid model-name '{}'!".format(model_name))
return
print("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
"model_{}_serialized_path".format(model_name))
print("Model deserialized from path '{}'".format(model_serialized_path))
camera = cv2.VideoCapture(0)
while True:
ret, frame = camera.read()
if not ret:
print("Failed to capture image!")
continue
frame = resize_image(frame, 400)
cv2.imshow("Webcam recording", frame)
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
classifier_model = joblib.load(model_serialized_path)
predicted_labels = classifier_model.predict(frame_flattened)
predicted_label = predicted_labels[0]
print("Predicted label = {}".format(predicted_label))
predicted_image = get_image_from_label(predicted_label)
predicted_image = resize_image(predicted_image, 200)
cv2.imshow("Prediction = '{}'".format(predicted_label),
predicted_image)
except Exception:
exception_traceback = traceback.format_exc()
print(
"Error while applying image transformation with the following exception trace:\n{}"
.format(exception_traceback))
cv2.waitKey(10000)
cv2.destroyAllWindows()
cv2.destroyAllWindows()
print "The program completed successfully !!"
def main():
model_name = sys.argv[1]
if model_name not in ['svm', 'logistic', 'knn']:
print("Invalid model-name '{}'!".format(model_name))
return
print("Using model '{}'...".format(model_name))
model_serialized_path = get_config(
'model_{}_serialized_path'.format(model_name))
print("Model deserialized from path '{}'".format(model_serialized_path))
testing_images_labels_path = get_config('testing_images_labels_path')
with open(testing_images_labels_path, 'r') as file:
lines = file.readlines()
for line in lines:
print("\n\n" + line.strip())
image_path, image_label = line.split()
frame = cv2.imread(image_path)
try:
frame = apply_image_transformation(frame)
frame_flattened = frame.flatten()
classifier_model = joblib.load(model_serialized_path)
predicted_labels = classifier_model.predict(frame_flattened)
predicted_label = predicted_labels[0]
print("Predicted label = {}".format(predicted_label))
engine = pyttsx.init()
engine.say("The predicted text is " + str(predicted_label))
engine.runAndWait()
engine.stop()
if image_label != predicted_label:
print("Incorrect prediction!!")
cv2.waitKey(5000)
except Exception:
exception_traceback = traceback.format_exc()
print("Error while applying image transformation on image path '{}' with the following exception trace:\n{}".format(
image_path, exception_traceback))
continue
cv2.waitKey()
cv2.destroyAllWindows()
print "The program completed successfully !!"
def main():
images_transformed_path = get_config('images_transformed_path')
images_transformed_path = get_config('images_transformed_path')
with open(images_transformed_path, 'w') as output_file:
writer = csv.writer(output_file, delimiter=',')
training_images_labels_path = get_config('training_images_labels_path')
with open(training_images_labels_path, 'r') as file:
lines = file.readlines()
images = []
labels = []
for line in lines:
#print("\n\n" + line.strip())
image_path, image_label = line.split()
# Read the input image.
frame = cv2.imread(image_path)
# `frame` is a HxW numpy ndarray of triplets (pixels), where H and W are
# the dimensions of the input image.
# cv2.imshow("Original", frame)
frame = apply_image_transformation(frame)
images.append(frame)
labels.append(image_label)
X_train, X_test, Y_train, Y_test = divide_data_train_test(
images, labels, 0.2)
X_train = np.array(X_train)
Y_train = np.array(Y_train)
Y_train = vec_translate(Y_train)
X_test = np.array(X_test)
Y_test = vec_translate(Y_test)
Y_test = np.array(Y_test)
X_train = X_train / 255.
X_test = X_test / 255.
Y_train = convert_to_one_hot(Y_train, 20).T
Y_test = convert_to_one_hot(Y_test, 20).T
#print(X_train.shape)
_, _, parameters = model(X_train, Y_train, X_test, Y_test)
print("Done!\n")
def main():
transformed_images_path = get_config('transformed_images_path')
os.makedirs(os.path.dirname(transformed_images_path), exist_ok=True)
with open(transformed_images_path, 'w') as output_file:
writer = csv.writer(output_file, delimiter=',')
training_images_labels_path = get_config('training_images_labels_path')
with open(training_images_labels_path, 'r') as file:
lines = file.readlines()
if show_progress_bar:
progress_bar = tqdm(total=len(lines),
desc='Transforming images',
unit=' pics')
for line in lines:
logger.debug("\n\n" + line.strip())
image_path, image_label = line.split()
# Read the input image.
frame = cv2.imread(image_path)
# `frame` is a HxW numpy ndarray of triplets (pixels),
# where H and W are the dimensions of the input image.
# cv2.imshow("Original", frame)
try:
if show_progress_bar:
progress_bar.update()
frame = apply_image_transformation(frame)
write_frame_to_file(frame, image_label, writer)
except Exception:
# Its normal to get errors on some images!!
exception_traceback = traceback.format_exc()
logger.error("Error applying image transformation to image "
"'{}'".format(image_path))
logger.debug(exception_traceback)
continue
# cv2.waitKey(1000)
cv2.destroyAllWindows()
if show_progress_bar:
progress_bar.close()
logger.info("The program completed successfully !!")