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 
import cv2

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

# 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 predict_on_video(video_in_filepath, video_out_filepath, detection_model, category_index):
    video_reader = cv2.VideoCapture(video_in_filepath)
    frame_h = int(video_reader.get(cv2.CAP_PROP_FRAME_HEIGHT))
    frame_w = int(video_reader.get(cv2.CAP_PROP_FRAME_WIDTH))
    fps = video_reader.get(cv2.CAP_PROP_FPS)
    
    video_writer = cv2.VideoWriter(
        video_out_filepath,
        cv2.VideoWriter_fourcc(*'mp4v'),
        fps,
        (frame_w, frame_h)
    )
    while True:
        ret, frame = video_reader.read()
        if not ret:
            break  # Break the loop if the video is finished
        
        processed_frame = predict(frame)
        processed_frame_np = np.array(processed_frame)
        video_writer.write(processed_frame_np)
                
    # Release camera and close windows
    video_reader.release()
    video_writer.release() 
    cv2.destroyAllWindows() 
    cv2.waitKey(1)        
    video_reader.release()
    video_writer.release()
    cv2.destroyAllWindows()
    cv2.waitKey(1)

# Function to process a video
def process_video(video_path):
    output_path = "output_video.mp4"  # Output path for the processed video
    predict_on_video(video_path, output_path, detection_model, category_index)
    return output_path

# Specify paths to example images
sample_images = [["00000031.jpg"], ["00000053.jpg"],
                 ["00000057.jpg"], ["00000078.jpg"], 
                 ["00000854.jpg"], ["00000995.jpg"],
                 ["00001052.jpg"],["00001444.jpg"],["00001452.jpg"]
                ]

REPO_ID = "jiawenchim/iti107model"
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')

tab1 = gr.Interface(fn=predict,
             inputs=gr.Image(type="pil"),
             outputs=gr.Image(type="pil"), 
             examples=sample_images, 
             title="Image - Object Detection (Battery and Dice)",
             description='Model used: SSD MobileNet V1 FPN 640x640.'
             )

tab2 = gr.Interface(
    fn=process_video,
    inputs=gr.File(label="Upload a video"),
    outputs=gr.File(label="output"),
    title='Video - Object Detection (Battery and Dice)',
    examples=[["Three Dice Trick.mp4"],["Look at the fork and battery-in power.mp4"]], 
    description='For video processing interface, I would like to endorse student 23B712M for his works. \n Model used: SSD MobileNet V1 FPN 640x640. \n Remarks: Running inference on Three Dice Tricks will take roughly 15-20mins'
)


iface = gr.TabbedInterface([tab1, tab2], tab_names  = ['Image','Video'], title='Battery and Dice detection')

iface.launch(share=True)