File size: 1,757 Bytes
77b834f
 
4cd0c8b
 
 
 
77b834f
4cd0c8b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
tags:
- background-removal
- Pytorch
- vision
---

# BRIA Background Removal v1.3

## Usage

```python
import os
import numpy as np
from skimage import io
from glob import glob
from tqdm import tqdm
import cv2
import torch.nn.functional as F
from torchvision.transforms.functional import normalize
from models import BriaRMBG

input_size=[1024,1024]
net=BriaRMBG()

model_path = "./model.pth"
im_path = "./example_image.jpg"
result_path = "."

if torch.cuda.is_available():
    net.load_state_dict(torch.load(model_path))
    net=net.cuda()
else:
    net.load_state_dict(torch.load(model_path,map_location="cpu"))
net.eval()    

# prepare input
im = io.imread(im_path)
if len(im.shape) < 3:
    im = im[:, :, np.newaxis]
im_size=im.shape[0:2]
im_tensor = torch.tensor(im, dtype=torch.float32).permute(2,0,1)
im_tensor = F.interpolate(torch.unsqueeze(im_tensor,0), size=input_size, mode='bilinear').type(torch.uint8)
image = torch.divide(im_tensor,255.0)
image = normalize(image,[0.5,0.5,0.5],[1.0,1.0,1.0])

if torch.cuda.is_available():
    image=image.cuda()

# inference 
result=net(image)

# post process
result = torch.squeeze(F.interpolate(result[0][0], size=im_size, mode='bilinear') ,0)
ma = torch.max(result)
mi = torch.min(result)
result = (result-mi)/(ma-mi)

# save result
im_name=im_path.split('/')[-1].split('.')[0]
im_array = (result*255).permute(1,2,0).cpu().data.numpy().astype(np.uint8)
cv2.imwrite(os.path.join(result_path, im_name+".png"), im_array)
```

## Training data
Bria-RMBG model was trained over 12000 high quality, high resolution images. 
All images were manualy labeled pixel-wise accuratly. The images belong to veriety of categories, the majority of them inclues people. 

## Qualitative Evaluation