app.py

import gradio as gr
from gradio_imageslider import ImageSlider
from loadimg import load_img
import spaces
from transformers import AutoModelForImageSegmentation
import torch
from torchvision import transforms
from PIL import Image

torch.set_float32_matmul_precision(["high", "highest"][0])

# Load BiRefNet model
birefnet = AutoModelForImageSegmentation.from_pretrained(
    "ZhengPeng7/BiRefNet", trust_remote_code=True
)
birefnet.to("cpu")

# Preprocessing pipeline
transform_image = transforms.Compose(
    [
        transforms.Resize((1024, 1024)),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
    ]
)


def process(image):
    """Segment person/body from image with transparent background"""
    image_size = image.size
    input_images = transform_image(image).unsqueeze(0).to("cpu")
    # Prediction
    with torch.no_grad():
        preds = birefnet(input_images)[-1].sigmoid().cpu()
    pred = preds[0].squeeze()
    pred_pil = transforms.ToPILImage()(pred)
    mask = pred_pil.resize(image_size)

    # Apply mask as alpha channel (transparent background)
    image = image.convert("RGBA")
    image.putalpha(mask)

    return image  # Transparent PNG


def fn(image):
    im = load_img(image, output_type="pil")
    im = im.convert("RGB")
    origin = im.copy()
    result = process(im)
    return result


def process_file(f):
    """Process uploaded file and save output as PNG with transparency"""
    name_path = f.rsplit(".", 1)[0] + ".png"
    im = load_img(f, output_type="pil")
    im = im.convert("RGB")
    transparent = process(im)
    transparent.save(name_path, "PNG")  # Save with transparency
    return name_path


# Gradio UI Components
slider1 = gr.Image()
slider2 = ImageSlider(label="birefnet", type="pil")
image = gr.Image(label="Upload an image")
image2 = gr.Image(label="Upload an image", type="filepath")
text = gr.Textbox(label="Paste an image URL")
png_file = gr.File(label="Output PNG file")

# Example image
chameleon = load_img("butterfly.jpg", output_type="pil")

# Tab for PNG output
tab3 = gr.Interface(
    process_file,
    inputs=image2,
    outputs=png_file,
    examples=["butterfly.jpg"],
    api_name="png",
)

# Main demo
demo = gr.TabbedInterface([tab3], ["PNG Output"], title="Body Extractor")

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