app.py

from flask import Flask, render_template, request, jsonify
import os
import cv2
import numpy as np
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.builders import model_builder
from object_detection.utils import config_util

app = Flask(__name__)

# Load model and label map
CUSTOM_MODEL_NAME = 'my_ssd_mobnet'
paths = {
    'CHECKPOINT_PATH': os.path.join('Tensorflow', 'workspace', 'models', CUSTOM_MODEL_NAME),
    'LABELMAP': os.path.join('Tensorflow', 'workspace', 'annotations', 'label_map.pbtxt')
}

configs = config_util.get_configs_from_pipeline_file(os.path.join(paths['CHECKPOINT_PATH'], 'pipeline.config'))
detection_model = model_builder.build(model_config=configs['model'], is_training=False)
ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
ckpt.restore(os.path.join(paths['CHECKPOINT_PATH'], 'ckpt-7')).expect_partial()
category_index = label_map_util.create_category_index_from_labelmap(paths['LABELMAP'])

# Define detection function
@tf.function
def detect_fn(image):
    image, shapes = detection_model.preprocess(image)
    prediction_dict = detection_model.predict(image, shapes)
    detections = detection_model.postprocess(prediction_dict, shapes)
    return detections

# Define route for object detection
@app.route('/detect', methods=['POST'])
def detect():
    # Get image file from request
    file = request.files['image']
    
    # Read image and convert to numpy array
    img = cv2.imdecode(np.frombuffer(file.read(), np.uint8), cv2.IMREAD_COLOR)
    image_np = np.array(img)
    
    # Perform object detection
    input_tensor = tf.convert_to_tensor(np.expand_dims(image_np, 0), dtype=tf.float32)
    detections = detect_fn(input_tensor)
    
    num_detections = int(detections.pop('num_detections'))
    detections = {key: value[0, :num_detections].numpy() for key, value in detections.items()}
    detections['num_detections'] = num_detections
    detections['detection_classes'] = detections['detection_classes'].astype(np.int64)
    
    label_id_offset = 1
    image_np_with_detections = image_np.copy()
    
    viz_utils.visualize_boxes_and_labels_on_image_array(
        image_np_with_detections,
        detections['detection_boxes'],
        detections['detection_classes'] + label_id_offset,
        detections['detection_scores'],
        category_index,
        use_normalized_coordinates=True,
        max_boxes_to_draw=10,
        min_score_thresh=.4,
        agnostic_mode=False
    )
    
    # Convert image back to byte stream
    ret, buffer = cv2.imencode('.jpg', image_np_with_detections)
    img_str = buffer.tobytes()
    
    return img_str

# Define index route
@app.route('/', methods=['GET'])
def index():
    return render_template('index.html')

if __name__ == "__main__":
    app.run(debug=True)