|
import tensorflow as tf |
|
import numpy as np |
|
from PIL import Image |
|
from glob import glob |
|
import pandas as pd |
|
from tensorflow.keras.preprocessing.image import img_to_array |
|
from huggingface_hub import from_pretrained_keras |
|
import gradio as gr |
|
|
|
model = from_pretrained_keras("keras-io/super-resolution") |
|
model.summary() |
|
|
|
def infer(image): |
|
|
|
nx=image.shape[0] |
|
ny=image.shape[1] |
|
|
|
img = Image.fromarray(image) |
|
|
|
|
|
ycbcr = img.convert("YCbCr") |
|
y, cb, cr = ycbcr.split() |
|
y = img_to_array(y) |
|
y = y.astype("float32") / 255.0 |
|
|
|
input = np.expand_dims(y, axis=0) |
|
out = model.predict(input) |
|
|
|
nxo = out.squeeze().shape[0] |
|
nyo = out.squeeze().shape[1] |
|
|
|
out_img_y = out[0] |
|
out_img_y *= 255.0 |
|
|
|
|
|
out_img_y = out_img_y.clip(0, 255) |
|
out_img_y = out_img_y.reshape((np.shape(out_img_y)[0], np.shape(out_img_y)[1])) |
|
out_img_y = Image.fromarray(np.uint8(out_img_y), mode="L") |
|
out_img_cb = cb.resize(out_img_y.size, Image.BICUBIC) |
|
out_img_cr = cr.resize(out_img_y.size, Image.BICUBIC) |
|
out_img = Image.merge("YCbCr", (out_img_y, out_img_cb, out_img_cr)).convert( |
|
"RGB" |
|
) |
|
|
|
out_img.save('output.png') |
|
|
|
out = {} |
|
out.update( {'input image size x': nx } ) |
|
out.update( {'output image size x': nxo } ) |
|
out.update( {'input image size y': ny } ) |
|
out.update( {'output image size y': nyo } ) |
|
|
|
return (pd.DataFrame(data=out.values(), index=out.keys()).transpose(), img,out_img, 'output.png') |
|
|
|
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/1609.05158' target='_blank'>Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network</a></p><center> <a href='https://keras.io/examples/vision/super_resolution_sub_pixel/' target='_blank'>Image Super-Resolution using an Efficient Sub-Pixel CNN</a></p>" |
|
|
|
examples= [[l] for l in glob('examples/tiles/*.jpg')] |
|
out1 = gr.outputs.Dataframe(label='Summary', headers=["Input X (px)", "Output X (px)", "Input Y (px)", "Output Y (px)"], type='pandas') |
|
out2 = gr.outputs.Image(label="Cropped input image", type='pil') |
|
out3 = gr.outputs.Image(label="Super-resolution x3 image", type='pil') |
|
out4 = gr.outputs.File(label='Click to download super-resolved image') |
|
|
|
iface = gr.Interface( |
|
fn=infer, |
|
title = " Satellite Super-resolution", |
|
description = "This space is a demo of Satellite image Super-Resolution using a Sub-Pixel Convolutional Neural Network", |
|
article = article, |
|
inputs=gr.inputs.Image(label="Input Image"), |
|
outputs=[out1,out2,out3,out4], |
|
examples=examples, |
|
).launch() |