app.py

"""
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()