app.py

import PIL.Image as Image
import gradio as gr
import cv2
import numpy as np

from ultralytics import ASSETS, YOLO

model = YOLO("best.pt")

def preprocess_image(img):
    # Convert to grayscale
    img_gray = img.convert("L")
    
    img_resized = img_gray.resize((640,640))
    
    # Apply CLAHE (Contrast Limited Adaptive Histogram Equalization)
    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
    img_clahe = clahe.apply(np.array(img_resized))
    
    
    return Image.fromarray(img_clahe)

def predict_image(img, conf_threshold, iou_threshold):
    # Preprocess the image
    img_preprocessed = preprocess_image(img)
    
    results = model.predict(
        source=img_preprocessed,
        conf=conf_threshold,
        iou=iou_threshold,
        show_labels=True,
        show_conf=True,
        imgsz=640,
    )

    for r in results:
        im_array = r.plot()
        im = Image.fromarray(im_array[..., ::-1])

    return im

iface = gr.Interface(
    fn=predict_image,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Slider(minimum=0, maximum=1, value=0.7, label="Confidence threshold"),
        gr.Slider(minimum=0, maximum=1, value=0.45, label="IoU threshold")
    ],
    outputs=gr.Image(type="pil", label="Result"))

if __name__ == '__main__':
    iface.launch()