app.py

from ultralytics import YOLO
from PIL import Image
import gradio as gr
from huggingface_hub import snapshot_download
import os
import cv2
import numpy as np

model_path = "best_int8_openvino_model"

def load_model(repo_id):
    download_dir = snapshot_download(repo_id)
    print(download_dir)
    path  = os.path.join(download_dir, "best_int8_openvino_model")
    print(path)
    detection_model = YOLO(path, task='detect')
    return detection_model


# Image prediction function
def predict_image(input_image):
    source = input_image
    result = detection_model.predict(source, conf=0.5, iou=0.6)
    img_bgr = result[0].plot()
    img_rgb = cv2.cvtColor(img_bgr, cv2.COLOR_BGR2RGB)  # Convert BGR to RGB
    output_image = Image.fromarray(img_rgb)  # Use the RGB image for output
    return output_image

# Video prediction function
def predict_video(input_video):
    # Get the original filename (without path)
    original_filename = os.path.basename(input_video.name)
    
    # Get the name without the extension and append '_detected'
    base_filename, _ = os.path.splitext(original_filename)
    output_video = base_filename + "_detected.mp4"
    
    # Read video file
    video_capture = cv2.VideoCapture(input_video.name)
    frames = []
    
    while True:
        ret, frame = video_capture.read()
        if not ret:
            break
        
        # Process each frame
        result = detection_model.predict(frame, conf=0.5, iou=0.6)
        img_bgr = result[0].plot()  # Get the frame with detected objects
        frames.append(img_bgr)  # Add the RGB frame to the list
    
    # Release video capture
    video_capture.release()

    # Convert frames to video
    fourcc = cv2.VideoWriter_fourcc(*'mp4v')  # Codec for mp4
    out = cv2.VideoWriter(output_video, fourcc, 30, (frames[0].shape[1], frames[0].shape[0]))
    
    for frame in frames:
        out.write(frame)  # Write each frame to video

    out.release()
    
    return output_video  # Return the path to the output video


REPO_ID = "AI-Pagoda/4673483T"
detection_model = load_model(REPO_ID)



# Create Gradio interface with tabs
with gr.Blocks() as app:
    with gr.Tabs():
        with gr.Tab("Image Detection"):
            gr.Interface(fn=predict_image,
                         inputs=gr.Image(type="pil", label="Upload Image"),
                         outputs=gr.Image(type="pil", label="Download Image"),
                         title="Image Object Detection",
                         description="Upload an image to detect Snake and Lizard.")

        with gr.Tab("Video Detection"):
            gr.Interface(fn=predict_video,
                         inputs=gr.File(label="Upload Video"),
                         outputs=gr.File(label="Download Video"),
                         title="Video Object Detection",
                         description="Upload an image to detect Snake and Lizard.")

app.launch()