File size: 3,758 Bytes
6c2a15f 4dc2bc3 6c2a15f 4dc2bc3 6c2a15f 4dc2bc3 6c2a15f 4dc2bc3 6c2a15f 4dc2bc3 6c2a15f 4dc2bc3 6c2a15f |
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 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 |
"""
gradio app.py
for semantic segmentation
"""
import os
import cv2
import gradio as gr
import numpy as np
from otfgt import mask2sbd
def gen_sbd(image, mask):
h, w = image.shape[:2]
if w > 1280 or h > 720:
resize_factor = max(w / 1280, h / 720)
h = int(h / resize_factor)
w = int(w / resize_factor)
image = cv2.resize(image, (w, h))
mask = cv2.resize(mask, (w, h), interpolation=cv2.INTER_NEAREST)
binary_labels = np.zeros((19, mask.shape[0], mask.shape[1]), dtype=np.uint8)
unique_labels = np.unique(mask)
for label in unique_labels:
binary_labels[label] = mask == label
sbd = mask2sbd(binary_labels, ignore_indices=[])
# remove the first channel (background)
sbd = sbd[1:]
unique_boundary_labels = np.unique(np.where(sbd == 1)[0])
value = [(sbd[x], ID2LABEL[x]) for x in unique_boundary_labels]
# change 0 to 255
mask[mask == 0] = 255
# reduce the entries by 1
mask -= 1
unique_labels = np.unique(mask)
# remove 254 (background) from unique_labels
unique_labels = unique_labels[unique_labels != 254]
value_segmentation = [(mask == x, ID2LABEL[x]) for x in unique_labels]
return (image, value), (image, value_segmentation)
HF_TOKEN = os.environ.get("HF_TOKEN", None)
ID2LABEL = { # id: label
0: "road",
1: "dirt",
2: "gravel",
3: "rock",
4: "grass",
5: "vegetation",
6: "tree",
7: "obstacle",
8: "animals",
9: "person",
10: "bicycle",
11: "vehicle",
12: "water",
13: "boat",
14: "building",
15: "roof",
16: "sky",
17: "drone",
}
input_1 = gr.Image(
image_mode="RGB",
type="numpy",
label="Image (RGB)",
)
input_2 = gr.Image(
image_mode="L",
type="numpy",
label="Segmentation Mask (Greyscale)",
)
INPUTS = [input_1, input_2]
output_1 = gr.AnnotatedImage(
label="Boundary Mask",
)
output_2 = gr.AnnotatedImage(
label="Segmentation Mask",
)
OUTPUTS = [output_1, output_2]
TITLE = "Semantic Boundary Generation"
DESCRIPTION = "Semantic Boundary Generation based on [paper](https://arxiv.org/pdf/2304.09427.pdf)."
theme = gr.themes.Monochrome(
primary_hue="indigo",
secondary_hue="blue",
neutral_hue="slate",
radius_size=gr.themes.sizes.radius_sm,
font=[
gr.themes.GoogleFont("Open Sans"),
"ui-sans-serif",
"system-ui",
"sans-serif",
],
)
cur_dir = os.path.dirname(os.path.abspath(__file__))
EXAMPLES = [
[
f"{cur_dir}/examples/aeroscape_1.jpg",
f"{cur_dir}/examples/aeroscape_1_mask.png",
],
[
f"{cur_dir}/examples/aeroscape_2.jpg",
f"{cur_dir}/examples/aeroscape_2_mask.png",
],
[
f"{cur_dir}/examples/floodnet_1.jpg",
f"{cur_dir}/examples/floodnet_1_mask.png",
],
[
f"{cur_dir}/examples/floodnet_2.jpg",
f"{cur_dir}/examples/floodnet_2_mask.png",
],
[
f"{cur_dir}/examples/floodnet_3.jpg",
f"{cur_dir}/examples/floodnet_3_mask.png",
],
[
f"{cur_dir}/examples/floodnet_4.jpg",
f"{cur_dir}/examples/floodnet_4_mask.png",
],
[
f"{cur_dir}/examples/floodnet_5.jpg",
f"{cur_dir}/examples/floodnet_5_mask.png",
],
[
f"{cur_dir}/examples/udd_1.jpg",
f"{cur_dir}/examples/udd_1_mask.png",
],
[
f"{cur_dir}/examples/udd_2.jpg",
f"{cur_dir}/examples/udd_2_mask.png",
],
]
demo = gr.Interface(
fn=gen_sbd,
inputs=INPUTS,
outputs=OUTPUTS,
title=TITLE,
description=DESCRIPTION,
live=False,
theme=theme,
allow_flagging="never",
cache_examples=True,
examples=EXAMPLES,
)
if __name__ == "__main__":
demo.launch()
|