def make_graph(self):
video_detections = load_pickle(
os.path.join(mydir, "data", "duke_test_seq_cam2_10.pck"))
ground_truth = load_pickle(
os.path.join(mydir, "data", "duke_test_seq_cam2_10_gt.pck"))
video_detections = [
(frame_idx, imread(os.path.join(mydir, frame)), detections)
for frame_idx, frame, detections in video_detections
]
return make_graph(video_detections, 60), ground_truth
def annotate(self):
for img_path in self.next():
img = imread(img_path)
color = (0, 255, 0) if self.left_to_annotate == 0 else (255, 0, 0)
img = cv2.putText(
img,
"%d: %s" % (self.current_idx, self.db.get(img_path, "-")),
(50, 100),
cv2.FONT_HERSHEY_SIMPLEX,
4,
color,
10,
)
self.view(img, pause=True)
print("Not annotated images: %d/%d" %
(self.left_to_annotate, len(self)))
def next(self):
headers = {}
while True:
l = self._fd.readline()
if not l:
raise StopIteration
if not l.strip() and headers:
break
if b':' in l:
i = l.index(b':')
headers[l[:i]] = l[i + 1:].strip()
data = self._fd.read(int(headers[b'Content-Length']))
img = imread(StringIO(data)).view(Frame)
img.index = self.fcnt
self.fcnt += 1
img.timestamp = img.systime = -1 # FIXME
return img
def next(self):
headers = {}
while True:
l = self._fd.readline()
if not l:
raise StopIteration
if not l.strip() and headers:
break
if ':' in l:
i = l.index(':')
headers[l[:i]] = l[i+1:].strip()
data = self._fd.read(int(headers['Content-Length']))
img = imread(StringIO(data)).view(Frame)
img.index = self.fcnt
self.fcnt += 1
img.timestamp = img.systime = -1 # FIXME
return img
def train(dataset, input_shape):
net = Net(input_shape)
optimizer = torch.optim.SGD(net.parameters(), lr=0.01)
loss = nn.CrossEntropyLoss()
net.train()
for i, (path, label) in enumerate(dataset.items_forever()):
# Read the input data and transform the label to index
img = imread(path)
img = cv2.resize(img, input_shape, interpolation=cv2.INTER_NEAREST)
data = torch.Tensor(img[:, :, 0].flatten())
target = torch.Tensor([0 if label == "open" else 1]).type(torch.long)
# Calculate gradients
optimizer.zero_grad()
output = net(data)
batch_loss = loss(output.view(1, -1), target)
batch_loss.backward() # Compute sum of gradients
optimizer.step()
print(
f"Iteration {i}: Output {output} vs expected {target} - loss: {batch_loss}"
)
def frame(self, scene, frame):
return imread(self._path(scene, "img1/%.6d.jpg" % frame))
self.on_mouse_motion(x, y)
self.clicking = False
def on_mouse_drag(self, x, y, dx, dy, buttons, modifiers):
DebugViewer.scroll[0] += dx
DebugViewer.scroll[1] += dy
self.clicking = False
def on_mouse_press(self, x, y, button, modifiers):
self.clicking = True
def on_mouse_release(self, x, y, button, modifiers):
if self.clicking:
print(DebugViewer.mouse_x, DebugViewer.mouse_y)
if __name__ == '__main__':
from vi3o import Video, view
import sys
if len(sys.argv) < 2:
print("Usage: python -mvi3o.debugview <video file>")
exit(-1)
for fn in sys.argv[1:]:
if fn.split('.')[-1].lower() in ('mkv', 'mjpg', 'avi', 'mov', 'mp4'):
for img in Video(sys.argv[1]):
view(img)
else:
from vi3o.image import imread, imshow
imshow(imread(fn))
self.on_mouse_motion(x, y)
self.clicking = False
def on_mouse_drag(self, x, y, dx, dy, buttons, modifiers):
DebugViewer.scroll[0] += dx
DebugViewer.scroll[1] += dy
self.clicking = False
def on_mouse_press(self, x, y, button, modifiers):
self.clicking = True
def on_mouse_release(self, x, y, button, modifiers):
if self.clicking:
print(DebugViewer.mouse_x, DebugViewer.mouse_y)
if __name__ == '__main__':
from vi3o import Video, view
import sys
if len(sys.argv) < 2:
print("Usage: python -mvi3o.debugview <video file>")
exit(-1)
for fn in sys.argv[1:]:
if fn.split('.')[-1] in ('mkv', 'mjpg'):
for img in Video(sys.argv[1]):
view(img)
else:
from vi3o.image import imread, imshow
imshow(imread(fn))
def frame(self, scene, frame):
fn = os.path.join(self._path(scene, 'sequences'), '%.7d.jpg' % frame)
return imread(fn)