app.py

import matplotlib.pyplot as plt
import numpy as np
from six import BytesIO
from PIL import Image
import tensorflow as tf
from object_detection.utils import label_map_util
from object_detection.utils import visualization_utils as viz_utils
from object_detection.utils import ops as utils_op
import tarfile
import wget 
import gradio as gr
from huggingface_hub import snapshot_download
import os 

PATH_TO_LABELS = 'data/label_map.pbtxt'   
category_index = label_map_util.create_category_index_from_labelmap(PATH_TO_LABELS, use_display_name=True)

def pil_image_as_numpy_array(pilimg):

    img_array = tf.keras.utils.img_to_array(pilimg)
    img_array = np.expand_dims(img_array, axis=0)
    return img_array
    
def load_image_into_numpy_array(path):
                                    
    image = None
    image_data = tf.io.gfile.GFile(path, 'rb').read()
    image = Image.open(BytesIO(image_data))
    return pil_image_as_numpy_array(image)            

def load_model():
    download_dir = snapshot_download(REPO_ID)
    saved_model_dir = os.path.join(download_dir, "saved_model")
    detection_model = tf.saved_model.load(saved_model_dir)
    return detection_model

def load_model2():
    wget.download("https://nyp-aicourse.s3-ap-southeast-1.amazonaws.com/pretrained-models/balloon_model.tar.gz")
    tarfile.open("balloon_model.tar.gz").extractall()
    model_dir = 'saved_model'    
    detection_model = tf.saved_model.load(str(model_dir))
    return detection_model    

# samples_folder = 'test_samples
# image_path = 'test_samples/sample_balloon.jpeg
# 

def predict(pilimg):

    image_np = pil_image_as_numpy_array(pilimg)
    return predict2(image_np)

def predict2(image_np):

    results = detection_model(image_np)

    # different object detection models have additional results
    result = {key:value.numpy() for key,value in results.items()}
    
    label_id_offset = 0
    image_np_with_detections = image_np.copy()

    viz_utils.visualize_boxes_and_labels_on_image_array(
        image_np_with_detections[0],
        result['detection_boxes'][0],
        (result['detection_classes'][0] + label_id_offset).astype(int),
        result['detection_scores'][0],
        category_index,
        use_normalized_coordinates=True,
        max_boxes_to_draw=200,
        min_score_thresh=.60,
        agnostic_mode=False,
        line_thickness=2)

    result_pil_img = tf.keras.utils.array_to_img(image_np_with_detections[0])
    
    return result_pil_img

def detect_video(video):
    # Create a video capture object
    cap = cv2.VideoCapture(video)

    # Process frames in a loop
    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break

        # Expand dimensions since model expects images to have shape: [1, None, None, 3]
        image_np_expanded = np.expand_dims(frame, axis=0)

        # Run inference
        output_dict = detection_model(image_np_expanded)

        # Extract detections
        boxes = output_dict['detection_boxes'][0].numpy()
        scores = output_dict['detection_scores'][0].numpy()
        classes = output_dict['detection_classes'][0].numpy().astype(np.int64)

        # Draw bounding boxes and labels
        image_np_with_detections = viz_utils.visualize_boxes_and_labels_on_image_array(
            frame,
            boxes,
            classes,
            scores,
            category_index,
            use_normalized_coordinates=True,
            max_boxes_to_draw=20,
            min_score_thresh=.5,
            agnostic_mode=False)

        # Yield the processed frame
        yield image_np_with_detections

    # Release resources
    cap.release()

REPO_ID = "apailang/mytfodmodel"
detection_model = load_model()
# pil_image = Image.open(image_path)
# image_arr = pil_image_as_numpy_array(pil_image)

# predicted_img = predict(image_arr)
# predicted_img.save('predicted.jpg')

gr.Interface(fn=predict,
             inputs=gr.Image(type="pil"),
             outputs=gr.Image(type="pil")
             ).launch(share=True)

# iface = gr.Blocks()

# iface.Image(fn=predict, inputs=gr.Image(type="pil"), label="Image Detection",outputs=gr.Image(type="pil"))
# iface.Video(fn=detect_video, inputs=gr.Video(type="mp4", live=True, label="Input Video"),outputs=gr.Video(type="mp4", label="Detected Video", live=True), label="Video Detection", interpretation="default")

# iface.launch(share=True)