app.py

import gradio as gr
import cv2
import numpy as np
from typing import Union, List
from pathlib import Path
from PIL import Image
import torch
from ultralytics.utils.plotting import save_one_box
from ultralytics.engine.results import Results
from ultralytics import YOLOWorld

# Function to resize images

def resize_images(images, scale_percent=50):
    resized_images = []
    for img in images:
        width = int(img.shape[1] * scale_percent / 100)
        height = int(img.shape[0] * scale_percent / 100)
        dim = (width, height)
        resized = cv2.resize(img, dim, interpolation=cv2.INTER_AREA)
        resized_images.append(resized)
    return resized_images

# Function to stitch images

def stitch_images(image_paths, scale_percent=50):
    images = [cv2.imread(path) for path in image_paths]
    resized_images = resize_images(images, scale_percent)
    stitcher = cv2.Stitcher_create()
    status, stitched_image = stitcher.stitch(resized_images)

    if status == cv2.Stitcher_OK:
        print("Stitching successful!")
        return stitched_image
    else:
        print(f"Stitching failed with status code: {status}")
        return None

# YOLO detection function

def yolo_detect(
    image: Union[str, Path, int, Image.Image, list, tuple, np.ndarray, torch.Tensor],
    classes: List[str],
) -> np.ndarray:
    detector = YOLOWorld("yolov8x-worldv2.pt")
    detector.set_classes(classes)
    results: List[Results] = detector.predict(image)

    boxes = []
    for result in results:
        boxes.append(
            save_one_box(result.cpu().boxes.xyxy, im=result.orig_img, save=False)
        )

    del detector
    return boxes[0]

# Function to predict and process image

def predict(image, offset_inches):
    try:
        drawer_img = yolo_detect(image, ["box"])
        shrunked_img = make_square(shrink_bbox(drawer_img, 0.8))
        return shrunked_img
    except:
        raise Exception("Unable to DETECT DRAWER, please take another picture with different magnification level!")

# Function to shrink bounding box

def shrink_bbox(image: np.ndarray, shrink_factor: float):
    height, width = image.shape[:2]
    center_x, center_y = width // 2, height // 2

    new_width = int(width * shrink_factor)
    new_height = int(height * shrink_factor)

    x1 = max(center_x - new_width // 2, 0)
    y1 = max(center_y - new_height // 2, 0)
    x2 = min(center_x + new_width // 2, width)
    y2 = min(center_y + new_height // 2, height)

    cropped_image = image[y1:y2, x1:x2]
    return cropped_image

# Function to make image square

def make_square(img: np.ndarray):
    height, width = img.shape[:2]
    max_dim = max(height, width)
    pad_height = (max_dim - height) // 2
    pad_width = (max_dim - width) // 2
    pad_height_extra = max_dim - height - 2 * pad_height
    pad_width_extra = max_dim - width - 2 * pad_width

    if len(img.shape) == 3:
        padded = np.pad(
            img,
            ((pad_height, pad_height + pad_height_extra),
             (pad_width, pad_width + pad_width_extra),
             (0, 0)),
            mode="edge"
        )
    else:
        padded = np.pad(
            img,
            ((pad_height, pad_height + pad_height_extra),
             (pad_width, pad_width + pad_width_extra)),
            mode="edge"
        )

    return padded

# Main image processing function

def process_image(image_paths, scale_percent=50, offset_inches=1):
    stitched_image = stitch_images(image_paths, scale_percent)

    if stitched_image is not None:
        try:
            stitched_image_rgb = cv2.cvtColor(stitched_image, cv2.COLOR_BGR2RGB)
            final_image = predict(stitched_image_rgb, offset_inches)
            return final_image
        except Exception as e:
            print(str(e))
            return stitched_image

# Gradio interface function

def gradio_stitch_and_detect(image_files):
    image_paths = [file.name for file in image_files]
    result_image = process_image(image_paths, scale_percent=50)

    if result_image is not None:
        result_image_rgb = cv2.cvtColor(result_image, cv2.COLOR_BGR2RGB)
        pil_image = Image.fromarray(result_image_rgb)
        pil_image.save("stitched_image.jpg", "JPEG")
        return pil_image, "stitched_image.jpg"

    return None, None

# Gradio interface
with gr.Blocks() as interface:
    gr.Markdown("<h1 style='color: #2196F3; text-align: center;'>Image Stitcher 🧵</h1>")
    gr.Markdown("<h3 style='color: #2196F3; text-align: center;'>=== Upload the images you want to stitch ===</h3>")

    image_upload = gr.Files(type="filepath", label="Upload Images")
    stitch_button = gr.Button("Stitch", variant="primary")
    stitched_image = gr.Image(type="pil", label="Stitched Image")
    download_button = gr.File(label="Download Stitched Image")

    stitch_button.click(gradio_stitch_and_detect, inputs=image_upload, outputs=[stitched_image, download_button])

interface.launch()