为什么使用 Tensorflows 对象检测 API 使用 Faster RCNN 进行预测的速度如此缓慢？

Question

我正在使用 Tensorflows 对象检测 api 来使用我自己的数据集来训练我自己的对象检测。我的问题是预测每张图像大约需要 25-30 秒。这包括读取图像和绘制边界框。我在具有 1 个核心的 NVIDIA K80 GPU 上运行。有什么指导可以加快速度吗？我基本上采用了来自 Object Detection api 的预测脚本。

这是我用来进行预测的两个函数：

def predict_image(TEST_IMAGE_PATHS, PATH_TO_CKPT, category_index, save_path):
    detection_graph = load_detection_graph(PATH_TO_CKPT)
    prediction_dict = defaultdict()
    start_time = time.time()
    for image_path in TEST_IMAGE_PATHS:
        toc = time.time()
        filename = image_path.split('/')[-1]
        image = Image.open(image_path)
        # the array based representation of the image will be used later in order to prepare the
        # result image with boxes and labels on it.
        image_np = load_image_into_numpy_array(image)
        # Expand dimensions since the model expects images to have shape: [1, None, None, 3]
        image_np_expanded = np.expand_dims(image_np, axis=0)
        # Actual detection.
        output_dict = run_inference_for_single_image(image_np, detection_graph, filename)
        # Visualization of the results of a detection.

        vis_util.visualize_boxes_and_labels_on_image_array(
          image_np,
          output_dict['detection_boxes'],
          output_dict['detection_classes'],
          output_dict['detection_scores'],
          category_index,
          instance_masks=output_dict.get('detection_masks'),
          use_normalized_coordinates=True,
          line_thickness=8)
        prediction_dict[filename] = output_dict

        tic = time.time()
        print('{0} saved in {1:.2f}sec'.format(filename, tic-toc))
    end_time = time.time()



def run_inference_for_single_image(image, graph, filename):
  with graph.as_default():
    with tf.Session() as sess:
      # Get handles to input and output tensors
      ops = tf.get_default_graph().get_operations()
      all_tensor_names = {output.name for op in ops for output in op.outputs}
      tensor_dict = {}
      for key in [
          'num_detections', 'detection_boxes', 'detection_scores',
          'detection_classes', 'detection_masks'
      ]:
        tensor_name = key + ':0'
        if tensor_name in all_tensor_names:
          tensor_dict[key] = tf.get_default_graph().get_tensor_by_name(
              tensor_name)
      if 'detection_masks' in tensor_dict:
        # The following processing is only for single image
        detection_boxes = tf.squeeze(tensor_dict['detection_boxes'], [0])
        detection_masks = tf.squeeze(tensor_dict['detection_masks'], [0])
        # Reframe is required to translate mask from box coordinates to image coordinates and fit the image size.
        real_num_detection = tf.cast(tensor_dict['num_detections'][0], tf.int32)
        detection_boxes = tf.slice(detection_boxes, [0, 0], [real_num_detection, -1])
        detection_masks = tf.slice(detection_masks, [0, 0, 0], [real_num_detection, -1, -1])
        detection_masks_reframed = utils_ops.reframe_box_masks_to_image_masks(
            detection_masks, detection_boxes, image.shape[0], image.shape[1])
        detection_masks_reframed = tf.cast(
            tf.greater(detection_masks_reframed, 0.5), tf.uint8)
        # Follow the convention by adding back the batch dimension
        tensor_dict['detection_masks'] = tf.expand_dims(
            detection_masks_reframed, 0)
      image_tensor = tf.get_default_graph().get_tensor_by_name('image_tensor:0')

      # Run inference
      output_dict = sess.run(tensor_dict,
                             feed_dict={image_tensor: np.expand_dims(image, 0)})

      # all outputs are float32 numpy arrays, so convert types as appropriate
      output_dict['filename'] = filename
      output_dict['num_detections'] = int(output_dict['num_detections'][0])
      output_dict['detection_classes'] = output_dict[
          'detection_classes'][0].astype(np.uint8)
      output_dict['detection_boxes'] = output_dict['detection_boxes'][0]
      output_dict['detection_scores'] = output_dict['detection_scores'][0]
      if 'detection_masks' in output_dict:
        output_dict['detection_masks'] = output_dict['detection_masks'][0]
  return output_dict

Answer 1

如果您在同一模型上进行多个预测，则可以通过加载检测图一次来减少预测时间，并且只为每个预测运行会话。第一次预测需要时间，但后续预测需要更少的时间。 但是如果对每个预测使用不同的模型，我没有找到任何出路。