def main():
model = M.yolo_darknet19(input_shape=(416, 416, 3), output_depth=5)
yolo_outputs = Y.yolo_head(model.output)
to_evaluate = Y.yolo_eval(yolo_outputs)
sess = K.get_session()
Y.predict(sess, data, to_evaluate)
def getupload():
global image, motor, helmet
if request.method == "POST":
# memory = request.files['memory']
# image, mot, hel = predict(memory.filename)
# motor+=mot
# helmet+=hel
# return redirect('/upload')
try:
memory = request.files['memory']
description = request.form['description']
print(memory.filename == "", description == "")
if memory.filename != "" and description != "":
print("One")
return render_template(
'upload.html',
errorMessage="Please either upload a photo or link a url")
# elif memory.filename != "" and description != "":
# return render_template('upload.html', errorMessage="Please either upload a photo or link a url.")
elif memory.filename != "":
print("Two")
print(memory.filename)
print("here1")
image, mot, hel = predict(memory.filename)
cv.imwrite("/static/new_output.jpg", image)
print("here2")
motor += mot
helmet += hel
print(memory.filename)
print("END")
elif description != "":
print("Three")
print(description)
image, stats = predict(description)
motor += stats[0]
helmet += stats[1]
print(memory.filename)
else:
print("Four")
return render_template(
'upload.html',
errorMessage="Please either upload a photo or link a url")
except:
print("EXCEPT")
return render_template(
'upload.html',
errorMessage="Please either upload a photo or link a url.")
return redirect("/results")
def test_train_yolo_framework(setup_model_config, setup_weights_file,
setup_dataset_folder, setup_image_and_its_boxes):
model_config = setup_model_config
pretrained_feature_file, weight_file = setup_weights_file
img_folder, ann_folder = setup_dataset_folder
# 1. Construct the model
yolo = create_yolo(model_config['architecture'], model_config['labels'],
model_config['input_size'], model_config['anchors'],
pretrained_feature_file)
# 2. warmup training
yolo.train(img_folder, ann_folder, 2, weight_file, 2, False, 1e-4, 10, 1,
2, img_folder, ann_folder)
# 3. Load the warmup trained weights
yolo.load_weights(weight_file)
# 4. actual training
yolo.train(img_folder, ann_folder, 12, weight_file, 2, False, 1e-4, 10, 1,
0, img_folder, ann_folder)
# 5. Load training image & predict objects
image, true_boxes = setup_image_and_its_boxes
boxes, probs = yolo.predict(image)
boxes = to_centroid(boxes)
assert len(boxes) == 2
assert len(probs) == 2
assert np.allclose(np.argmax(probs, axis=1), [0, 3])
for box, true_box in zip(boxes, true_boxes):
iou = centroid_box_iou(box, true_box)
assert iou > 0.4
def detect(img_path): # tar in en bild och kör modellen över den.
image = img_path # om man bar vill testa med en bild byt mot cv2.imread(img_path)
boxes, probs = yolo.predict(image, float(DEFAULT_THRESHOLD))
image = draw_scaled_boxes(image, boxes, probs, labels,1)
return image
def get_digits_boxes_img(image):
im_size = np.shape(image)
boxes, probs = yolo.predict(image, float(threshold))
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
probs = np.max(probs, axis=1) if len(probs) > 0 else []
# 4. save detection result
image = draw_scaled_boxes(image, boxes, probs, config['model']['labels'])
centers = []
for count in range(len(boxes)):
box = boxes[count]
x = (((box[0] + box[2] / 2) / im_size[0]))
y = (((box[1] + box[3] / 2) / im_size[1]))
centers.append([x, y])
centers = np.array(centers)
jersey_numbers = []
if len(centers) > 1:
cluster.fit_predict(centers)
cluster_labels = cluster.labels_
clusters = np.unique(cluster_labels)
conf_score = []
for c_id in clusters:
g_centers = centers[list(cluster_labels == c_id)]
g_labels = labels[list(cluster_labels == c_id)]
g_probs = probs[list(cluster_labels == c_id)]
center_x = [center[0] for center in g_centers]
g_labels = g_labels[np.argsort(center_x)]
number = int((''.join(str(label) for label in g_labels)))
jersey_numbers.append(number)
conf_score.append(np.mean(g_probs))
else:
jersey_numbers = list(labels)
conf_score = [np.mean(probs)]
# print("{}-boxes are detected".format(len(boxes)))
# print("jersey numbers detected are {}".format(jersey_numbers))
return jersey_numbers, conf_score
def predict(real_time=False):
sess = K.get_session()
X_test, Y_test = utils.load_data('data.h5', 0, 1)
model = M.yolo_darknet19(input_shape=(416, 416, 3), output_depth=5)
model.load_weights('darknet19_weights_full_66.h5')
yolo_outputs = Y.yolo_head(model.output)
to_evaluate = Y.yolo_eval(yolo_outputs)
p, x, y = Y.predict(sess, X_test, to_evaluate, model)
m = p.shape[0]
if not real_time:
for i in range(m):
print('Image_' + str(i) + ' : with probability of ' + str(p[i]) +
'ball coords are (' + str(x[i]) + '; ' + str(y[i]) + ')')
utils.draw_result(X_test[i, ...], x[i], y[i])
else:
pass
def test_predict(setup_image_and_its_boxes, setup_outputs, setup_config):
# 1. Given
image, true_boxes = setup_image_and_its_boxes
model_config = setup_config
desired_boxes, desired_probs = setup_outputs
# 2. When run
# 1. Construct the model
yolo = create_yolo(model_config['architecture'], model_config['labels'],
model_config['input_size'], model_config['anchors'])
yolo.load_weights(os.path.join(TEST_SAMPLE_DIR, "mobile_288_weights.h5"))
boxes, probs = yolo.predict(image)
boxes = to_centroid(boxes)
assert len(boxes) == 2
assert len(probs) == 2
for box, true_box in zip(boxes, true_boxes):
iou = centroid_box_iou(box, true_box)
assert iou > 0.5
config['train']['valid_annot_folder'],
config['train']['valid_image_folder'],
config['model']['labels'],
is_only_detect=config['train']['is_only_detect'])
n_true_positives = 0
n_truth = 0
n_pred = 0
for i in range(len(annotations)):
img_path = annotations.fname(i)
img_fname = os.path.basename(img_path)
image = cv2.imread(img_path)
true_boxes = annotations.boxes(i)
true_labels = annotations.code_labels(i)
boxes, probs = yolo.predict(image, float(args.threshold))
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
# 4. save detection result
image = draw_scaled_boxes(image, boxes, probs,
config['model']['labels'])
output_path = os.path.join(write_dname, os.path.split(img_fname)[-1])
cv2.imwrite(output_path, image)
print("{}-boxes are detected. {} saved.".format(
len(boxes), output_path))
n_true_positives += count_true_positives(boxes, true_boxes, labels,
true_labels)
n_truth += len(true_boxes)
n_pred += len(boxes)
elif spotifyObject_1.currently_playing(
)["is_playing"]: # spela om musiken är pausad
spotifyObject_1.pause_playback()
print("pause playback")
time.sleep(0.5)
postion = []
cap = cv2.VideoCapture(0)
Connect_to_token()
creat_list()
while True:
_, frame = cap.read()
#startar detektering
boxes, probs = yolo.predict(frame, config["model"]["DEFAULT_THRESHOLD"])
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
#sparar detekterings resultat
frame = draw_scaled_boxes(frame, boxes, probs, config['model']['labels'])
label_list = config['model']['labels']
cv2.imshow("imageo", frame)
cv2.waitKey(1)
if 0 in labels: #om en tummeup hittas
try:
Add_to_list()
except:
input(
" No active device found ... press ENTER when you are back on track"
)
import numpy as np
import cv2
from yolo import predict
from PIL import Image
cap = cv2.VideoCapture(0)
cv2.waitKey(1000)
print("Start.")
i=0
key=1
while(i<100):
ret, frame = cap.read()
if cv2.waitKey(200) == 27:
break
i+=1
frame=predict(frame)
cv2.imshow("PIC",frame)
cv2.imwrite(f"/Users/Dennis/Desktop/hand genture project/pics/image-{i}.png",frame)
cap.release()
def get_digits_boxes(img, show_img=False, save_img=False):
img_path = img_path.replace('\\', '/')
# img_path = img_path.replace('\','/')
image = cv2.imread(img_path)
im_size = np.shape(image)
boxes, probs = yolo.predict(image, float(threshold))
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
probs = np.max(probs, axis=1) if len(probs) > 0 else []
# 4. save detection result
image = draw_scaled_boxes(image, boxes, probs, config['model']['labels'])
centers = []
for count in range(len(boxes)):
box = boxes[count]
x = (((box[0] + box[2] / 2) / im_size[0]))
y = (((box[1] + box[3] / 2) / im_size[1]))
centers.append([x, y])
centers = np.array(centers)
jersey_numbers = []
if len(centers) > 1:
cluster.fit_predict(centers)
cluster_labels = cluster.labels_
clusters = np.unique(cluster_labels)
conf_score = []
for c_id in clusters:
g_centers = centers[list(cluster_labels == c_id)]
g_labels = labels[list(cluster_labels == c_id)]
g_probs = probs[list(cluster_labels == c_id)]
center_x = [center[0] for center in g_centers]
g_labels = g_labels[np.argsort(center_x)]
number = int((''.join(str(label) for label in g_labels)))
jersey_numbers.append(number)
conf_score.append(np.mean(g_probs))
else:
jersey_numbers = labels
conf_score = np.mean(probs)
numbers_string = ''
for number in jersey_numbers:
numbers_string = numbers_string + str(number) + ' '
cv2.putText(image, 'Jersey Number: %s' % numbers_string, (25, 20),
cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 0), 2)
# labels=labels[np.argsort(center_x)]
# number = int((''.join(str(label) for label in labels)))
if save_img:
output_path = os.path.join(
ntpath.dirname(img_path) + '/' +
(img_path.split("/")[-1]).split('.')[0] + '_out.jpg')
cv2.imwrite(output_path, image)
if show_img:
cv2.imshow("jersey number detected image", image)
cv2.waitKey(5000)
cv2.destroyAllWindows()
print("{}-boxes are detected".format(len(boxes)))
print("jersey numbers detected are {}".format(jersey_numbers))
return jersey_numbers, conf_score, labels, boxes
# -*- coding: utf-8 -*-
"""image_od
Automatically generated by Colaboratory.
Original file is located at
https://colab.research.google.com/drive/1bYWenJ9BGhpIs7HgljiKhs0lX9_hzv73
# **Import libraries**
"""
import argparse
from yolo import predict
from od_utils import *
"""**construct the argument parse and parse the arguments**"""
ap = argparse.ArgumentParser()
ap.add_argument("-i", "--image", required=True, help='input image')
#ap.add_argument('-y', '--yolo', required=True, help='base path to YOLO model')
ap.add_argument('-c', '--confidence', type=float, default=0.6, help='minimum prob. to filter weak detections')
ap.add_argument('-t', '--threshold', type=float, default=0.5, help='threshold when applying non-maxima suppression')
args = vars(ap.parse_args())
"""**Run the graph on an image**"""
sess = K.get_session()
out_scores, out_boxes, out_classes = predict(sess, args)
annotations = parse_annotation(config['train']['valid_annot_folder'],
config['train']['valid_image_folder'],
config['model']['labels'],
is_only_detect=config['train']['is_only_detect'])
n_true_positives = 0
n_truth = 0
n_pred = 0
for i in range(len(annotations)):
img_path = annotations.fname(i)
img_fname = os.path.basename(img_path)
image = cv2.imread(img_path)
true_boxes = annotations.boxes(i)
true_labels = annotations.code_labels(i)
boxes, probs = yolo.predict(image, float(args.threshold))
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
# 4. save detection result
image = draw_scaled_boxes(image, boxes, probs, config['model']['labels'])
output_path = os.path.join(write_dname, os.path.split(img_fname)[-1])
cv2.imwrite(output_path, image)
print("{}-boxes are detected. {} saved.".format(len(boxes), output_path))
n_true_positives += count_true_positives(boxes, true_boxes, labels, true_labels)
n_truth += len(true_boxes)
n_pred += len(boxes)
print(calc_score(n_true_positives, n_truth, n_pred))
config['train']['valid_annot_folder'],
config['train']['valid_image_folder'],
config['model']['labels'],
is_only_detect=config['train']['is_only_detect'])
# n_true_positives = 0
# n_truth = 0
# n_pred = 0
# for i in range(len(annotations)):
for filename in os.listdir('datasets/' + evaluation_object +
'/images_test'):
img_path = os.path.join('datasets', evaluation_object, 'images_test',
filename)
img_fname = filename
image = cv2.imread(img_path)
boxes, probs = yolo.predict(image, float(DEFAULT_THRESHOLD))
labels = np.argmax(probs, axis=1) if len(probs) > 0 else []
# 4. save detection result
image = draw_scaled_boxes(image, boxes, probs,
config['model']['labels'])
output_path = os.path.join(write_dname, os.path.split(img_fname)[-1])
label_list = config['model']['labels']
# cv2.imwrite(output_path, image)
print("{}-boxes are detected. {} saved.".format(
len(boxes), output_path))
if len(probs) > 0:
create_ann(filename, image, boxes, labels, label_list)
cv2.imwrite(output_path, image)