app.py

import os
import cv2
import numpy as np
import importlib.util
from PIL import Image
import gradio as gr
from common_detection import perform_detection, resize_image

# Function to load the TensorFlow Lite model and labels
def load_model_and_labels(model_dir):
    pkg = importlib.util.find_spec('tflite_runtime')
    if pkg:
        from tflite_runtime.interpreter import Interpreter
    else:
        from tensorflow.lite.python.interpreter import Interpreter

    PATH_TO_CKPT = os.path.join(model_dir, 'detect.tflite')
    PATH_TO_LABELS = os.path.join(model_dir, 'labelmap.txt')

    with open(PATH_TO_LABELS, 'r') as f:
        labels = [line.strip() for line in f.readlines()]

    if labels[0] == '???':
        del(labels[0])

    interpreter = Interpreter(model_path=PATH_TO_CKPT)
    interpreter.allocate_tensors()

    input_details = interpreter.get_input_details()
    output_details = interpreter.get_output_details()
    height = input_details[0]['shape'][1]
    width = input_details[0]['shape'][2]
    floating_model = (input_details[0]['dtype'] == np.float32)

    return interpreter, labels, input_details, output_details, height, width, floating_model

# Load models
models = {
    "Multi-class model": "model",
    "Empty class": "model_2",
    "Misalignment class": "model_3"
}

# Function to perform image detection
def detect_image(model_choice, input_image):
    model_dir = models[model_choice]
    interpreter, labels, input_details, output_details, height, width, floating_model = load_model_and_labels(model_dir)
    image = np.array(input_image)
    resized_image = resize_image(image, size=640)
    result_image = perform_detection(resized_image, interpreter, labels, input_details, output_details, height, width, floating_model)
    return Image.fromarray(result_image)

# Function to perform video detection
def detect_video(model_choice, input_video):
    model_dir = models[model_choice]
    interpreter, labels, input_details, output_details, height, width, floating_model = load_model_and_labels(model_dir)
    cap = cv2.VideoCapture(input_video)
    frames = []

    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break

        resized_frame = resize_image(frame, size=640)
        result_frame = perform_detection(resized_frame, interpreter, labels, input_details, output_details, height, width, floating_model)
        frames.append(result_frame)

    cap.release()

    if not frames:
        raise ValueError("No frames were read from the video.")

    height, width, layers = frames[0].shape
    size = (width, height)
    output_video_path = "result_" + os.path.basename(input_video)
    out = cv2.VideoWriter(output_video_path, cv2.VideoWriter_fourcc(*'mp4v'), 15, size)

    for frame in frames:
        out.write(frame)

    out.release()

    return output_video_path

app = gr.Blocks()

with app:
    gr.Markdown("## Object Detection using TensorFlow Lite Models")
    with gr.Row():
        model_choice = gr.Dropdown(label="Select Model", choices=["Multi-class model", "Empty class", "Misalignment class"])
        with gr.Tab("Image Detection"):
            image_input = gr.Image(type="pil", label="Upload an image")
            image_output = gr.Image(type="pil", label="Detection Result")
            gr.Button("Submit Image").click(fn=detect_image, inputs=[model_choice, image_input], outputs=image_output)
        with gr.Tab("Video Detection"):
            video_input = gr.Video(label="Upload a video")
            video_output = gr.Video(label="Detection Result")
            gr.Button("Submit Video").click(fn=detect_video, inputs=[model_choice, video_input], outputs=video_output)

app.launch(share=True)