def init_yolo():
global yolo
config = "models/cross-hands.cfg"
weights = "models/cross-hands.weights"
if not (os.path.isfile(config) and os.path.isfile(weights)):
subprocess.call("./models/download-some-models.sh", shell=True)
yolo = YOLO(config, weights, ["hand"])
yolo.size = int(256)
yolo.confidence = float(0.3)
yolo = YOLO("models/cross-hands.cfg", "models/cross-hands.weights",
["hand"])
while True:
timecounter = 0
starter = 0
ix = 0
iy = 0
ixx = 0
iyy = 0
t = 0
arrows = cv2.imread('images/sofa.jpeg')
sofa = cv2.imread('images/sofa.jpeg')
yolo.size = int(args.size)
yolo.confidence = float(args.confidence)
print("starting webcam...")
cv2.namedWindow("preview")
# cv2.setWindowProperty ("preview", cv2.WND_PROP_FULLSCREEN, cv2.WINDOW_FULLSCREEN)
try:
vc = cv2.VideoCapture(0)
if vc.isOpened(): # try to get the first frame
rval, frame = vc.read()
else:
rval = False
while rval:
while frame.shape[0] < arrows.shape[0] or frame.shape[
def scrollV(cap, network, device, size, confidence):
if network == "normal":
print("loading yolo...")
yolo = YOLO("models/cross-hands.cfg", "models/cross-hands.weights",
["hand"])
elif network == "prn":
print("loading yolo-tiny-prn...")
yolo = YOLO("models/cross-hands-tiny-prn.cfg",
"models/cross-hands-tiny-prn.weights", ["hand"])
else:
print("loading yolo-tiny...")
yolo = YOLO("models/cross-hands-tiny.cfg",
"models/cross-hands-tiny.weights", ["hand"])
yolo.size = size
yolo.confidence = confidence
cnt = 0
curr = 0
prev = 0
exit = 0
rval, frame = cap.read()
while True:
width, height, inference_time, results = yolo.inference(frame)
if len(results) == 1:
exit = 0
cnt += 1
id, name, confidence, x, y, w, h = results[0]
cx = x + (w // 2)
cy = y + (h // 2)
if cnt <= 5:
curr = cy
color = (0, 255, 255)
cv2.circle(frame, (cx, cy), 10, color, -1)
#print("Cy: ", cy)
if cnt % 10 == 0 and cnt > 5:
prev = curr
curr = cy
#print("Prev: ",prev)
#print("Curr: ", curr)
clicks = prev - curr
#print(clicks)
#if clicks>30 and clicks<170:
clicks = clicks // 2
if abs(clicks) > 10:
pyautogui.scroll(clicks)
else:
exit += 1
if exit > 50:
print(exit)
break
cv2.imshow("preview", frame)
rval, frame = cap.read()
key = cv2.waitKey(1)
if key == 27: # exit on ESC
break
cv2.destroyWindow("preview")
import argparse
import cv2
from yolo import YOLO
yolo = YOLO("models/cross-hands.cfg", "models/cross-hands.weights", ["hand"])
yolo.size = 416
yolo.confidence = 0.2
print("Iniciando webcam...")
cv2.namedWindow("Camera")
vc = cv2.VideoCapture(0)
if vc.isOpened():
rval, frame = vc.read()
else:
rval = False
while rval:
width, height, inference_time, results = yolo.inference(frame)
for detection in results:
id, name, confidence, x, y, w, h = detection
cx = x + (w / 2)
cy = y + (h / 2)
color = (0, 255, 255)
cv2.rectangle(frame, (x, y), (x + w, y + h), color, 2)
text = "%s (%s)" % (name, round(confidence, 2))
cv2.putText(frame, text, (x, y - 5), cv2.FONT_HERSHEY_SIMPLEX,
0.5, color, 2)
if network == "normal":
print("loading yolo...")
yolo = YOLO("models/cross-hands.cfg", "models/cross-hands.weights", ["hand"])
elif network == "prn":
print("loading yolo-tiny-prn...")
yolo = YOLO("models/cross-hands-tiny-prn.cfg", "models/cross-hands-tiny-prn.weights", ["hand"])
elif network == "v4-tiny":
print("loading yolov4-tiny-prn...")
yolo = YOLO("models/cross-hands-yolov4-tiny.cfg", "models/cross-hands-yolov4-tiny.weights", ["hand"])
else:
print("loading yolo-tiny...")
yolo = YOLO("models/cross-hands-tiny.cfg", "models/cross-hands-tiny.weights", ["hand"])
yolo.size = int(size)
yolo.confidence = float(confidence)
@app.route("/hand_detection")
def hand_detection():
image_path = request.args.get("image_path")
print(image_path)
mat = cv2.imread(image_path)
width, height, inference_time, results = yolo.inference(mat)
print("%s seconds: %s classes found!" % (round(inference_time, 2), len(results)))
if len(results) < 1:
return json.dumps({"nums_of_hand": 0})
# Khatna Bold
# Adapted from: https://github.com/cansik/yolo-hand-detection/tree/master/
import argparse
import cv2
import numpy as np
from yolo import YOLO
from PIL import Image
import utils
yolo = YOLO("models/cross-hands-tiny-prn.cfg",
"models/cross-hands-tiny-prn.weights", ["hand"])
yolo.size = int(512)
yolo.confidence = float(0.2)
# main Loop
print("starting webcam...")
vc = cv2.VideoCapture(0)
detecting = True
code = ''
text = '---'
correct_count = 0
while True:
_, frame = vc.read()
frame = np.flip(frame, axis=1)
H, W, _ = frame.shape
ref = int(min(H, W) / 3)
window = None
if detecting:
width, height, inference_time, results = yolo.inference(frame)
def reading_video(filename):
ap = argparse.ArgumentParser()
ap.add_argument('-n',
'--network',
default="normal",
help='Network Type: normal / tiny / prn / v4-tiny')
ap.add_argument('-d', '--device', default=0, help='Device to use')
ap.add_argument('-v',
'--videos',
default="videos",
help='Path to videos or video file')
ap.add_argument('-s', '--size', default=416, help='Size for yolo')
ap.add_argument('-c',
'--confidence',
default=0.2,
help='Confidence for yolo')
ap.add_argument("-f",
"--fff",
help="a dummy argument to fool ipython",
default="1")
args = ap.parse_args()
if args.network == "normal":
print("loading yolo...")
yolo = YOLO("models/cross-hands.cfg", "models/cross-hands.weights",
["hand"])
elif args.network == "prn":
print("loading yolo-tiny-prn...")
yolo = YOLO("models/cross-hands-tiny-prn.cfg",
"models/cross-hands-tiny-prn.weights", ["hand"])
elif args.network == "v4-tiny":
print("loading yolov4-tiny-prn...")
yolo = YOLO("models/cross-hands-yolov4-tiny.cfg",
"models/cross-hands-yolov4-tiny.weights", ["hand"])
else:
print("loading yolo-tiny...")
yolo = YOLO("models/cross-hands-tiny.cfg",
"models/cross-hands-tiny.weights", ["hand"])
yolo.size = int(args.size)
yolo.confidence = float(args.confidence)
# opening a window called preview
cv2.namedWindow("preview")
# to open and capture frames from video
vc = cv2.VideoCapture(filename)
if vc.isOpened(): # try to get the first frame
rval, frame = vc.read()
# to get the first frame
else:
# some error causes the video to not open
rval = False
while (vc.isOpened()):
# Applying YOLO on the frames
width, height, inference_time, results = yolo.inference(frame)
for detection in results:
id, name, confidence, x, y, w, h = detection
cx = x + (w / 2)
cy = y + (h / 2)
# draw a bounding box rectangle and label on the image
color = (0, 255, 255)
cv2.rectangle(frame, (x, y), (x + w, y + h), color, 2)
text = "%s (%s)" % (name, round(confidence, 2))
# put a text on the detected hand with the confidence ratio
cv2.putText(frame, text, (x, y - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.5,
color, 2)
cv2.imshow("preview", frame)
rval, frame = vc.read()
# to close the window we need to click on the ESC button
key = cv2.waitKey(20)
if key == 27: # exit on ESC
break
cv2.destroyWindow("preview")
vc.release()