700 KiB
700 KiB
In [4]:
import cv2
from pathlib import Path
import numpy as np
# from PIL import Image
import torch
from torchvision.io.video import read_video
import matplotlib.pyplot as plt
from torchvision.utils import draw_bounding_boxes
from torchvision.transforms.functional import to_pil_image
from torchvision.models.detection import retinanet_resnet50_fpn_v2, RetinaNet_ResNet50_FPN_V2_Weights
In [5]:
source = Path('../DATASETS/VIRAT_subset_0102x')
videos = source.glob('*.mp4')
homography = list(source.glob('*img2world.txt'))[0]
H = np.loadtxt(homography, delimiter=',')
The homography matrix helps to transform points from image space to a flat world plane. The README_homography.txt from VIRAT describes:
Roughly estimated 3-by-3 homographies are included for convenience. Each homography H provides a mapping from image coordinate to scene-dependent world coordinate.
[xw,yw,zw]' = H*[xi,yi,1]'xi: horizontal axis on image with left top corner as origin, increases right. yi: vertical axis on image with left top corner as origin, increases downward.
xw/zw: world x coordinate yw/zw: world y coordiante
In [6]:
# H.dot(np.array([20,300, 1]))
In [7]:
video_path = list(videos)[0]
video_path = Path("../DATASETS/VIRAT_subset_0102x/VIRAT_S_010200_00_000060_000218.mp4")
In [8]:
video_path
Out[8]:
In [9]:
weights = RetinaNet_ResNet50_FPN_V2_Weights.DEFAULT
model = retinanet_resnet50_fpn_v2(weights=weights, score_thresh=0.35)
# Put the model in inference mode
model.eval()
# Get the transforms for the model's weights
preprocess = weights.transforms()
In [10]:
# hub.set_dir()
In [11]:
video = cv2.VideoCapture(str(video_path))
The score_thresh argument defines the threshold at which an object is detected as an object of a class. Intuitively, it's the confidence threshold, and we won't classify an object to belong to a class if the model is less than 35% confident that it belongs to a class.
The result from a single prediction coming from model(batch) looks like:
{'boxes': tensor([[5.7001e+02, 2.5786e+02, 6.3138e+02, 3.6970e+02],
[5.0109e+02, 2.4508e+02, 5.5308e+02, 3.4852e+02],
[3.4096e+02, 2.7015e+02, 3.6156e+02, 3.1857e+02],
[5.0219e-01, 3.7588e+02, 9.7911e+01, 7.2000e+02],
[3.4096e+02, 2.7015e+02, 3.6156e+02, 3.1857e+02],
[8.3241e+01, 5.8410e+02, 1.7502e+02, 7.1743e+02]]),
'scores': tensor([0.8525, 0.6491, 0.5985, 0.4999, 0.3753, 0.3746]),
'labels': tensor([64, 64, 1, 64, 18, 86])}
In [37]:
# TODO make into loop
%matplotlib inline
import pylab as pl
from IPython import display
i=0
while True:
ret, frame = video.read()
i+=1
if not ret:
print("Can't receive frame (stream end?). Exiting ...")
break
t = torch.from_numpy(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
t.shape
# image = image[np.newaxis, :]
t = t.permute(2, 0, 1)
t.shape
batch = [preprocess(t)]
# no_grad can be used on inference, should be slightly faster
with torch.no_grad():
predictions = model(batch)
prediction = predictions[0] # we feed only one frame at the once
mask = prediction['labels'] == 1 # if we want more than one: np.isin(prediction['labels'], [1,86])
scores = prediction['scores'][mask]
labels = prediction['labels'][mask]
boxes = prediction['boxes'][mask]
# TODO: introduce confidence and NMS supression: https://github.com/cfotache/pytorch_objectdetecttrack/blob/master/PyTorch_Object_Tracking.ipynb
# (which I _think_ we better do after filtering)
# alternatively look at Soft-NMS https://towardsdatascience.com/non-maximum-suppression-nms-93ce178e177c
labels = [weights.meta["categories"][i] for i in labels]
box = draw_bounding_boxes(t, boxes=boxes,
labels=labels,
colors="cyan",
width=2,
font_size=30,
font='Arial')
im = to_pil_image(box.detach())
display.display(im, f"frame {i}")
print(prediction)
display.clear_output(wait=True)
break # for now
# pl.clf()
# # pl.plot(pl.randn(100))
# pl.figure(figsize=(24,50))
# # fig.axes[0].imshow(img)
# pl.imshow(im)
# display.display(pl.gcf(), f"frame {i}")
# display.clear_output(wait=True)
# time.sleep(1.0)
# fig, ax = plt.subplots(figsize=(16, 12))
# ax.imshow(im)
# plt.show()
In [36]:
prediction
Out[36]:
In [21]:
prediction['labels'] == 1
Out[21]:
In [23]:
prediction['boxes'][prediction['labels'] == 1]
prediction['scores'][prediction['labels'] == 1]
Out[23]:
In [35]:
Out[35]:
Now with SORT tracking¶
Using a sort implementation originally by Alex Bewley, but adapted by Chris Fotache. For an example implementation, see his notebook.
In [39]:
from sort_cfotache import Sort
mot_tracker = Sort()
In [38]:
# TODO make into loop
%matplotlib inline
import pylab as pl
from IPython import display
i=0
while True:
ret, frame = video.read()
i+=1
if not ret:
print("Can't receive frame (stream end?). Exiting ...")
break
t = torch.from_numpy(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
t.shape
# image = image[np.newaxis, :]
t = t.permute(2, 0, 1)
t.shape
batch = [preprocess(t)]
# no_grad can be used on inference, should be slightly faster
with torch.no_grad():
predictions = model(batch)
prediction = predictions[0] # we feed only one frame at the once
mask = prediction['labels'] == 1 # if we want more than one: np.isin(prediction['labels'], [1,86])
scores = prediction['scores'][mask]
labels = prediction['labels'][mask]
boxes = prediction['boxes'][mask]
# TODO: introduce confidence and NMS supression: https://github.com/cfotache/pytorch_objectdetecttrack/blob/master/PyTorch_Object_Tracking.ipynb
# (which I _think_ we better do after filtering)
# alternatively look at Soft-NMS https://towardsdatascience.com/non-maximum-suppression-nms-93ce178e177c
labels = [weights.meta["categories"][i] for i in labels]
box = draw_bounding_boxes(t, boxes=boxes,
labels=labels,
colors="cyan",
width=2,
font_size=30,
font='Arial')
im = to_pil_image(box.detach())
display.display(im, f"frame {i}")
print(prediction)
display.clear_output(wait=True)
break # for now
# pl.clf()
# # pl.plot(pl.randn(100))
# pl.figure(figsize=(24,50))
# # fig.axes[0].imshow(img)
# pl.imshow(im)
# display.display(pl.gcf(), f"frame {i}")
# display.clear_output(wait=True)
# time.sleep(1.0)
# fig, ax = plt.subplots(figsize=(16, 12))
# ax.imshow(im)
# plt.show()
In [ ]: