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 |