app.py

import gradio as gr
from matplotlib import gridspec
import matplotlib.pyplot as plt
import numpy as np
from PIL import Image
import torch
from transformers import AutoImageProcessor, AutoModelForSemanticSegmentation

MODEL_ID = "tobiasc/segformer-b0-finetuned-segments-sidewalk"
processor = AutoImageProcessor.from_pretrained(MODEL_ID)
model = AutoModelForSemanticSegmentation.from_pretrained(MODEL_ID)

def ade_palette():
    """ADE20K palette that maps each class to RGB values."""
    return [
        [0, 0, 0],  # 0: unlabeled
        [120, 120, 120],  # 1: flat-road (회색)
        [244, 35, 232],  # 2: flat-sidewalk (분홍)
        [107, 142, 35],  # 3: flat-crosswalk (녹색)
        [70, 130, 180],  # 4: flat-cyclinglane (하늘색)
        [255, 0, 0],  # 5: flat-parkingdriveway (빨강)
        [0, 0, 142],  # 6: flat-railtrack (진청)
        [220, 20, 60],  # 7: flat-curb (진홍)
        [220, 220, 0],  # 8: human-person (노랑)
        [119, 11, 32],  # 9: human-rider (적갈)
        [0, 0, 230],  # 10: vehicle-car (파랑)
        [0, 0, 70],  # 11: vehicle-truck (남색)
        [0, 60, 100],  # 12: vehicle-bus (청록)
        [0, 80, 100],  # 13: vehicle-tramtrain
        [0, 0, 110],  # 14: vehicle-motorcycle
        [111, 74, 0],  # 15: vehicle-bicycle
        [51, 51, 0],  # 16: vehicle-caravan
        [81, 0, 81],  # 17: vehicle-cartrailer
        [70, 70, 70],  # 18: construction-building (진회색)
        [150, 100, 100],  # 19: construction-door
        [190, 153, 153],  # 20: construction-wall
        [153, 153, 153],  # 21: construction-fenceguardrail
        [102, 102, 156],  # 22: construction-bridge
        [128, 64, 128],  # 23: construction-tunnel (보라)
        [64, 170, 64],  # 24: construction-stairs
        [250, 170, 30],  # 25: object-pole (주황)
        [255, 255, 0],  # 26: object-trafficsign
        [152, 251, 152],  # 27: object-trafficlight
        [31, 119, 180],  # 28: nature-vegetation (초록)
        [174, 199, 232],  # 29: nature-terrain (연청)
        [255, 127, 14],  # 30: sky (연주황)
        [140, 86, 75],  # 31: void-ground
        [148, 103, 189],  # 32: void-dynamic
        [227, 119, 194],  # 33: void-static
        [188, 189, 34]  # 34: void-unclear
    ]

labels_list = []
with open("labels.txt", "r", encoding="utf-8") as fp:
    for line in fp:
        labels_list.append(line.rstrip("\n"))

colormap = np.asarray(ade_palette(), dtype=np.uint8)

def label_to_color_image(label):
    if label.ndim != 2:
        raise ValueError("Expect 2-D input label")
    if np.max(label) >= len(colormap):
        raise ValueError("label value too large.")
    return colormap[label]

def draw_plot(pred_img, seg_np):
    fig = plt.figure(figsize=(20, 15))
    grid_spec = gridspec.GridSpec(1, 2, width_ratios=[6, 1])

    plt.subplot(grid_spec[0])
    plt.imshow(pred_img)
    plt.axis('off')

    LABEL_NAMES = np.asarray(labels_list)
    FULL_LABEL_MAP = np.arange(len(LABEL_NAMES)).reshape(len(LABEL_NAMES), 1)
    FULL_COLOR_MAP = label_to_color_image(FULL_LABEL_MAP)

    unique_labels = np.unique(seg_np.astype("uint8"))
    ax = plt.subplot(grid_spec[1])
    plt.imshow(FULL_COLOR_MAP[unique_labels].astype(np.uint8), interpolation="nearest")
    ax.yaxis.tick_right()
    plt.yticks(range(len(unique_labels)), LABEL_NAMES[unique_labels])
    plt.xticks([], [])
    ax.tick_params(width=0.0, labelsize=25)
    return fig

def run_inference(input_img):
    # input: numpy array from gradio -> PIL
    img = Image.fromarray(input_img.astype(np.uint8)) if isinstance(input_img, np.ndarray) else input_img
    if img.mode != "RGB":
        img = img.convert("RGB")

    inputs = processor(images=img, return_tensors="pt")
    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits  # (1, C, h/4, w/4)

    # resize to original
    upsampled = torch.nn.functional.interpolate(
        logits, size=img.size[::-1], mode="bilinear", align_corners=False
    )
    seg = upsampled.argmax(dim=1)[0].cpu().numpy().astype(np.uint8)  # (H,W)

    # colorize & overlay
    color_seg = colormap[seg]                                # (H,W,3)
    pred_img = (np.array(img) * 0.5 + color_seg * 0.5).astype(np.uint8)

    fig = draw_plot(pred_img, seg)
    return fig

CUSTOM_CSS = """
:root { 
    --body-background-fill: #E0F7FA !important;
}
"""
demo = gr.Interface(
    fn=run_inference,
    inputs=gr.Image(type="numpy", label="Input Image"),
    outputs=gr.Plot(label="Overlay + Legend"),
    #theme="mono",
    css=CUSTOM_CSS,
    examples=[
        "image1.jpg",
        "image2.jpg",
        "image3.jpg",
        "image4.jpg",
        "image5.jpg"
    ],
    title="⚡ ML Homework3: 보도블럭 세그멘테이션",
    description="이미지를 업로드하면 AI가 보도블럭, 도로, 사람 등을 자동으로 탐지합니다.",
    flagging_mode="never",
    cache_examples=False,
)

if __name__ == "__main__":
    demo.launch()