app.py

import torch
from transformers import pipeline

from PIL import Image

import cv2
import numpy as np

from random import choice
import io

detector50 = pipeline(model="facebook/detr-resnet-50")

detector101 = pipeline(model="facebook/detr-resnet-101")


import gradio as gr

COLORS = ["#ff7f7f", "#ff7fbf", "#ff7fff", "#bf7fff",
            "#7f7fff", "#7fbfff", "#7fffff", "#7fffbf",
            "#7fff7f", "#bfff7f", "#ffff7f", "#ffbf7f"]

fdic = {
    "family" : "Impact",
    "style" : "italic",
    "size" : 15,
    "color" : "yellow",
    "weight" : "bold"
}


def get_figure(in_pil_img, in_results):
    # Convert PIL image to OpenCV format
    img_cv2 = np.array(in_pil_img)
    img_cv2 = cv2.cvtColor(img_cv2, cv2.COLOR_RGB2BGR)
    
    for prediction in in_results:
        selected_color = choice(COLORS)
        color = tuple(int(selected_color[i:i+2], 16) for i in (1, 3, 5))  # Convert hex color to RGB tuple

        x, y = prediction['box']['xmin'], prediction['box']['ymin']
        w, h = prediction['box']['xmax'] - prediction['box']['xmin'], prediction['box']['ymax'] - prediction['box']['ymin']

        # Draw bounding box using OpenCV
        img_cv2 = cv2.rectangle(img_cv2, (x, y), (x+w, y+h), color, 2)
        text = f"{prediction['label']}: {round(prediction['score']*100, 1)}%"
        img_cv2 = cv2.putText(img_cv2, text, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, color, 2)

    # Convert back to PIL format
    img_pil = Image.fromarray(cv2.cvtColor(img_cv2, cv2.COLOR_BGR2RGB))
    return img_pil


def infer(model, in_pil_img):

    results = None
    if model == "detr-resnet-101":
        results = detector101(in_pil_img)
    else:
        results = detector50(in_pil_img)

    output_pil_img = get_figure(in_pil_img, results)

    output_pil_img.save("output.jpg")
     
    return output_pil_img


with gr.Blocks(title="DETR Object Detection using openCV",
                    css=".gradio-container {background:lightyellow;}"
               ) as demo:
    #sample_index = gr.State([])

    gr.HTML("""<div style="font-family:'Times New Roman', 'Serif'; font-size:16pt; font-weight:bold; text-align:center; color:royalblue;">ObjecTron🪄</div>""")
    gr.HTML("""<div style="font-family:'Times New Roman', 'Serif'; font-size:16pt; font-weight:bold; text-align:center; color:royalblue;">
            A object detection app using OpenCV, Huggingface-transformers, detr-resnet and Gradio </div>""")
    gr.HTML("""<h4 style="color:navy;">1. Select a model.</h4>""")

    model = gr.Radio(["detr-resnet-50", "detr-resnet-101"], value="detr-resnet-50", label="Model name")

    gr.HTML("""<br/>""")
    gr.HTML("""<h4 style="color:navy;">2-a. Select an example below</h4>""")
    gr.HTML("""<h4 style="color:navy;">2-b. Or upload an image by clicking on the canvas.</h4>""")

    with gr.Row():
        input_image = gr.Image(label="Input image", type="pil")
        output_image = gr.Image(label="Output image with predicted instances", type="pil")

    gr.Examples(['samples/god.jpg','samples/road.jpg','samples/cats.jpg','samples/detectron.png','samples/dogandcat.jpg'], inputs=input_image)

    gr.HTML("""<br/>""")
    gr.HTML("""<h4 style="color:navy;">3. Then, click the button below to predict and see the magic!!!</h4>""")

    send_btn = gr.Button("Expecto Patronum 🪄")
    send_btn.click(fn=infer, inputs=[model, input_image], outputs=[output_image])

    gr.HTML("""<br/>""")
    gr.HTML("""<h4 style="color:navy;">Reference</h4>""")
    gr.HTML("""<ul>""")
    gr.HTML("""<li><a href="https://colab.research.google.com/github/facebookresearch/detr/blob/colab/notebooks/detr_attention.ipynb" target="_blank">Hands-on tutorial for DETR by facebookresearch</a>""")
    gr.HTML("""</ul>""")


#demo.queue()
demo.launch(debug=True)