Spaces:
Sleeping
Sleeping
File size: 3,646 Bytes
7862f4b 66760e9 7862f4b 04693a8 7862f4b 29e2a42 7862f4b 66760e9 7862f4b b089b57 7862f4b b089b57 7862f4b b089b57 7862f4b b089b57 7862f4b b089b57 7d9b815 29e2a42 7d9b815 7862f4b b089b57 7862f4b b089b57 7862f4b b089b57 7d9b815 b089b57 145955f 7862f4b 48fbcef 77d41cc 54bb111 145955f 7862f4b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 |
import cv2
import gradio as gr
import imutils
import numpy as np
import torch
from PIL import Image
import torchvision.transforms as transforms
def load_model():
# CNN3D Layer's architecture
cnndata = CNNData(in_dim = 1,
n_f =[32,48],
kernel_size=[(5,5,5), (3,3,3)],
activations=[nn.ReLU(),nn.ReLU()],
bns = [True, True],
dropouts = [0, 0],
paddings = [(0,0,0),(0,0,0)],
strides = [(2,2,2),(2,2,2)])
# Feedforward layer's architecture
lindata = LinData(in_dim = conv3D_output_size(cnndata, [30, 256, 342]),
hidden_layers= [256,256,1],
activations=[nn.ReLU(),nn.ReLU(),None],
bns=[False,False,False],
dropouts =[0.2, 0, 0])
# combined architecture
args = NetData(cnndata, lindata)
# weight file
#weight_file = 'cnn3d_epoch_300.pt'
# CNN3D model
# device = 'cuda' if torch.cuda.is_available() else 'cpu'
device = torch.device('cpu')
cnn3d = CNN3D(args).to(device)
#cnn3d.load_state_dict(torch.load(os.path.join(base_path,'weights',weight_file), map_location=device))
cnn3d.eval()
#print(cnn3d)
return cnn3d
def parse_video(video_file):
"""A utility to parse the input videos.
Reference: https://pyimagesearch.com/2018/11/12/yolo-object-detection-with-opencv/
"""
vs = cv2.VideoCapture(video_file)
# try to determine the total number of frames in the video file
try:
prop = (
cv2.cv.CV_CAP_PROP_FRAME_COUNT
if imutils.is_cv2()
else cv2.CAP_PROP_FRAME_COUNT
)
total = int(vs.get(prop))
print("[INFO] {} total frames in video".format(total))
# an error occurred while trying to determine the total
# number of frames in the video file
except:
print("[INFO] could not determine # of frames in video")
print("[INFO] no approx. completion time can be provided")
total = -1
frames = []
# loop over frames from the video file stream
while True:
# read the next frame from the file
(grabbed, frame) = vs.read()
if frame is not None:
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
frames.append(frame)
# if the frame was not grabbed, then we have reached the end
# of the stream
if not grabbed:
break
return frames
def pil_parser(video_file):
model = load_model()
# cv2 parsing
dummy_frames = parse_video(video_file)
X = []
frames = np.arange(2,62,2)
use_transform : transforms.Compose =transforms.Compose([transforms.Resize([256, 342]),
transforms.ToTensor(),
transforms.Normalize(mean=[0.5], std=[0.5])])
for i in frames:
image = Image.fromarray(dummy_frames[i]).convert('L')
if use_transform is not None:
image = use_transform(image)
else:
image = transforms.ToTensor()(image)
X.append(image)
X = torch.stack(X, dim=1).unsqueeze(0)
return 'shape is : '+ str(X.shape)
example_list=[
["2350.mp4"],
["2300.mp4"],
]
gr.Interface(
fn=pil_parser,
inputs=gr.Video(label="Upload a video file"),
outputs="text",
examples=example_list,
title="Viscosity Regression From Video Data",
description=(
"Gradio demo for Video Regression"
),
allow_flagging='never',
).launch() |