如何使用 pytorch 对图像进行双三次上采样？

Question

I have png image. I want to upsample it using bicubic interpolation. I found this function in pytorch:

nn.functional.upsample(mode = "bicubic)

https://pytorch.org/docs/stable/generated/torch.nn.Upsample.html

But how to apply to my png image? Should I turn my image into some torch tesnor? I just haven't found any example of using this function exactly on png image

Answer 1

You can do this

import torch 
import torchvision.transforms as transforms
from PIL import Image


t = transforms.ToTensor()
img = Image.open("Table.png")

b = torch.nn.functional.upsample(t(img).unsqueeze(0),(500,400),mode = "bicubic")

you can also apply Bicubic using Image

img = Image.open("Table.png")
re = img.resize((400, 400),Image.BICUBIC)