app.py

import gradio as gr
from ultralytics import YOLO
from PIL import Image
from ultralytics.utils.plotting import Annotator, colors
import glob

# Load model and data
model = YOLO('Dental_model.pt')
pic_files = glob.glob('*.jpg')
names = model.model.names
cloud_sdk = '<iframe src="https://cloudhand-sdk-lmpc.vercel.app/" width="100%" height="600px" style="border:none;"></iframe>'

def detect_objects(image):
    image1 = image.copy()
    results = model.predict(image)
    classes = results[0].boxes.cls.cpu().tolist()
    boxes = results[0].boxes.xyxy.cpu()
    annotator = Annotator(image, line_width=3)
    annotator1 = Annotator(image1, line_width=3)
    
    for box, cls in zip(boxes, classes):
        annotator.box_label(box, label=names[int(cls)], color=colors(int(cls)))
        annotator1.box_label(box, label=None, color=colors(int(cls)))

    return Image.fromarray(annotator.result()), Image.fromarray(annotator1.result())

# Gradio Blocks App
with gr.Blocks() as demo:
    gr.Markdown("## Dental Analysis")
    gr.Markdown("Analyze your Dental XRAY image with our AI object Detection model")

    with gr.Row():
        with gr.Column():
            image_input = gr.Image(type="pil", label="Upload Image")
            run_button = gr.Button("Run Detection")
            example_images = gr.Examples(
                examples=pic_files,
                inputs=image_input,
                label="Examples"
            )
        with gr.Column():
            image_output_1 = gr.Image(type="pil", label="Dental Analysis")
            image_output_2 = gr.Image(type="pil", label="Without Labels")
            html_output = gr.HTML(cloud_sdk)

    run_button.click(fn=detect_objects, 
                     inputs=image_input, 
                     outputs=[image_output_1, image_output_2])

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