pytorch transform数据处理转c++
python推理代码转c++ sdk过程遇到pytorch数据处理的转换
1.python代码
import torch
from PIL import Image
from torchvision import transforms
data_transform = transforms.Compose(
[transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])
img = Image.open(img_path)
img = data_transform(img)2.transforms.Resize(256)
Parameters
size (sequence or int) –
Desired output size. If size is a sequence like (h, w), output size will be matched to this. If size is an int, smaller edge of the image will be matched to this number. i.e, if height > width, then image will be rescaled to (size * height / width, size).
3.transforms.ToTensor()
Convert a PIL Image or numpy.ndarray to tensor. This transform does not support torchscript.
Converts a PIL Image or numpy.ndarray (H x W x C) in the range [0, 255] to a torch.FloatTensor of shape (C x H x W) in the range [0.0, 1.0] if the PIL Image belongs to one of the modes (L, LA, P, I, F, RGB, YCbCr, RGBA, CMYK, 1) or if the numpy.ndarray has dtype = np.uint8
cv::Mat ClsSixPrivate::processImage(cv::Mat &img) {
int inW = img.cols;
int inH = img.rows;
cv::Mat croped_image;
if (inW > inH)
{
int newWidth = 256 * inW / inH;
cv::resize(img, img, cv::Size(newWidth, 256), 0, 0, cv::INTER_LINEAR);
croped_image = img(cv::Rect((newWidth - 224) / 2, 16, 224, 224)).clone();
}
else {
int newHeight= 256 * inH / inW;
cv::resize(img, img, cv::Size(256, newHeight), 0, 0, cv::INTER_LINEAR);
croped_image = img(cv::Rect(16, (newHeight - 224) / 2, 224, 224)).clone();
}
std::vector<float> mean_value{ 0.485, 0.456,0.406 };
std::vector<float> std_value{ 0.229, 0.224, 0.225 };
cv::Mat dst;
std::vector<cv::Mat> rgbChannels(3);
cv::split(croped_image, rgbChannels);
for (auto i = 0; i < rgbChannels.size(); i++)
{
rgbChannels[i].convertTo(rgbChannels[i], CV_32FC1, 1.0 / (std_value[i] * 255.0), (0.0 - mean_value[i]) / std_value[i]);
}
cv::merge(rgbChannels, dst);
return dst;
}总结
以上为个人经验，希望能给大家一个参考，也希望大家多多支持编程网。
