File size: 906 Bytes
d3092e8 |
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 |
import numpy as np
import utils
from keras.models import load_model
import matplotlib.pyplot as plt
import gradio as gr
model = load_model("building_segmentation_model.keras")
def segment(img):
mask = utils.predict(model, img)
mask = mask[..., 0]
# Set masked regions to blue with 70% transparency
blue_color = np.array([255, 0, 0], dtype=np.uint8) # Red color
transparency = 0.7
masked_image = img.copy()
masked_image[mask == 1] = masked_image[mask == 1] * (1 - transparency) + transparency * blue_color
return masked_image
demo = gr.Interface(
fn = segment,
inputs=[gr.Image()],
outputs=[gr.Image()],
examples=[
"images/22678945_15.png",
"images/22828930_15.png",
"images/23429080_15.png",
"images/23729035_15.png",
"images/24179065_15.png",
]
)
if __name__ == "__main__":
demo.launch() |