def __init__(self):
self.name = "YOLO-Recogniser"
self.log("Loading weights...")
self.model = lightnet.load("yolo")
self.log("Created")
def load_resources():
"""Loads in YoloV2 model
Returns:
model: YoloV2 model
"""
app.logger.debug('Loading yolo model')
return lightnet.load(Yolo.MODEL)
def __init__(self):
self.cap = cv2.VideoCapture(camera_port)
self.model = lightnet.load('yolo')
self.objects = []
self.es_dict = {
"bottle": "una botella",
'keyboard': 'un teclado',
'diningtable': 'una mesa',
'cup': 'una taza',
'laptop': 'un portatil'
}
def __init__(self):
self.cap = cv2.VideoCapture(camera_port)
self.cap.set(cv2.CAP_PROP_FRAME_WIDTH, 1280)
self.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 720)
self.cap.set(cv2.CAP_PROP_FPS, 10)
self.model = lightnet.load('yolo')
self.objects = []
self.es_dict = { "bottle" : "una botella" , \
'keyboard':'un teclado' , \
'diningtable':'una mesa' , \
'cup':'una taza' , \
'laptop':'un portatil' , \
'wine glass': 'una copa' }
def __init__(self, pPedidos, capture):
# Constant values
self.videoCapture = capture
self.cameraPort = 0
self.img_file_path = "data/img.jpg"
self.boxes = 0
self.beverage_list = ['bottle', 'vase', 'cup', 'wine glass']
self.beverage_dict = { 'person':["persona"], \
'bottle':["gaseosa", "cerveza"] , \
'cell phone':["celular"], \
'vase':["agua"] , \
'cup':["cafe","café","tinto","te","té"] , \
'wine glass':["copa de vino","vino"]}
# Variables
self.pedidos = pPedidos
self.available_drinks = []
self.model = lightnet.load('yolo')
def __init__(self):
self.txt_line = ""
self.cap = cv2.VideoCapture(camera_port)
self.cap.set(cv2.CAP_PROP_FRAME_WIDTH, 1280)
self.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 720)
self.cap.set(cv2.CAP_PROP_FPS, 30)
self.bridge = CvBridge()
self.model = lightnet.load('yolo')
self.objects = []
self.es_dict = { 'bottle' : 'botella' , \
'keyboard':'teclado' , \
'diningtable':'mesa' , \
'cup':'taza' , \
'vase':'vaso' , \
'laptop':'portatil' , \
'wine glass': 'copa' }
self.frame = None
rospy.wait_for_service("sIA_take_picture")
self.takePicture = rospy.ServiceProxy("sIA_take_picture", TakePicture)
self.boxes = []
self.dists = []
#!/usr/bin/python
# OBJECT RECOGNITION
# by tHE iNCREDIBLE mACHINE
#
# A nice script to test in
import lightnet
print("Loading weights...")
model = lightnet.load("yolo")
print("Loading image...")
image = lightnet.Image.from_bytes(open('/VirtualShare/dog.jpg', 'rb').read())
print("Classifying...")
boxes = model(image)
print(boxes)
print("Done.")
def plot_trajectory(_c, im):
for t in _c.keys():
x = []
y = []
for i, j in _c[t]:
x.append(i)
y.append(j)
plt.plot(x, y, zorder=1)
plt.imshow(im, zorder=0)
plt.axis([0, 238, 158, 0])
plt.show()
model = lightnet.load('yolo')
start = time.time()
collect = {}
# for dir in os.scandir(base_path):
prev_arr = []
for i in range(200):
fname = str(i+1).zfill(3) + '.tif'
image, size, im_p = get_jpg(fname)
tmp = model(image)
result = []
for i, e in enumerate(tmp):
result.append([i, e[-1][0], e[-1][1]])
def __init__(self):
self.model = lightnet.load('yolo')
def image_test(dataset, lightnet_model, source=None, api=None, exclude=None):
"""
Test Prodigy's image annotation interface with a YOLOv2 model loaded
via LightNet. Requires the LightNet library to be installed. The recipe
will find objects in the images, and create a task for each object.
"""
log("RECIPE: Starting recipe image.test", locals())
try:
import lightnet
except ImportError:
prints("Can't find LightNet", "In order to use this recipe, you "
"need to have LightNet installed (currently compatible with "
"Mac and Linux): pip install lightnet. For more details, see: "
"https://github.com/explosion/lightnet",
error=True,
exits=1)
def get_image_stream(model, stream, thresh=0.5):
for eg in stream:
if not eg['image'].startswith('data'):
msg = "Expected base64-encoded data URI, but got: '{}'."
raise ValueError(msg.format(eg['image'][:100]))
image = lightnet.Image.from_bytes(b64_uri_to_bytes(eg['image']))
boxes = [b for b in model(image, thresh=thresh) if b[2] >= thresh]
eg['width'] = image.width
eg['height'] = image.height
eg['spans'] = [get_span(box) for box in boxes]
for i in range(len(eg['spans'])):
task = copy.deepcopy(eg)
task['spans'][i]['hidden'] = False
task = set_hashes(task, overwrite=True)
score = task['spans'][i]['score']
task['score'] = score
yield task
def get_span(box, hidden=True):
class_id, name, prob, abs_points = box
name = str(name, 'utf8') if not isinstance(name, str) else name
x, y, w, h = abs_points
rel_points = [[x - w / 2, y - h / 2], [x - w / 2, y + h / 2],
[x + w / 2, y + h / 2], [x + w / 2, y - h / 2]]
return {
'score': prob,
'label': name,
'label_id': class_id,
'points': rel_points,
'center': [abs_points[0], abs_points[1]],
'hidden': hidden
}
model = lightnet.load(lightnet_model)
log("RECIPE: Loaded LightNet model {}".format(lightnet_model))
stream = get_stream(source, api=api, loader='images', input_key='image')
stream = fetch_images(stream)
def free_lighnet(ctrl):
nonlocal model
del model
return {
'view_id': 'image',
'dataset': dataset,
'stream': get_image_stream(model, stream),
'exclude': exclude,
'on_exit': free_lighnet
}
