对于相同图像大小，mask-rcnn的多个图像推理比快速rcnn的运行慢约10倍

Question

我使用自己的自定义数据集成功地重新训练了mask-rcnn和更快的rcnn模型，并且我想对多个图像进行推理。我使用以下代码修改了演示中的单个图像推断功能。如果我使用重新训练的fast-rcnn resnet101，则会得到以下结果 如果使用重新训练的mask-rcnn resnet101，则会得到以下结果 如果我使用fast-rcnn inception-resnet运行以下命令 和以下与mask-rcnn inception-resnet 所有图像的分辨率均为1024x768。请帮助这是否是正确的行为。谢谢

以下功能是我在演示中修改的功能

id=foo

以下是运行该功能的一段代码

def run_inference_for_multiple_images(images, graph):
  with graph.as_default():
    with tf.Session() as sess:
        output_dict_array = []
        dict_time = []
        for image in images:
            # 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:
                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
            start = time.time()
            output_dict = sess.run(tensor_dict,
                                   feed_dict={image_tensor: np.expand_dims(image, 0)})
            end = time.time()
            print('inference time : {}'.format(end-start))

            # all outputs are float32 numpy arrays, so convert types as appropriate
            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]

            output_dict_array.append(output_dict)
            dict_time.append(end-start)
return output_dict_array, dict_time

Answer 1

我发现了问题，并且以下功能似乎可以正常工作。每个图像的平均推断时间从大约3-4秒减少到0.3-0.4秒（使用resnet50特征提取器）。但是，使用此功能时必须小心，因为使用批量图像的假设是所有图像的尺寸必须相同。因此，当批处理中的图像之一具有不同大小时，将引发错误。虽然我自己还没有证实。

def run_inference_for_multiple_images(images, graph):
    with graph.as_default():
        with tf.Session() as sess:
            output_dict_array = []
            dict_time = []
            # 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:
                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, images[0].shape[0], images[0].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')
            for image in images:
                # Run inference
                start = time.time()
                output_dict = sess.run(tensor_dict, feed_dict={image_tensor: np.expand_dims(image, 0)})
                end = time.time()
                print('inference time : {}'.format(end - start))

                # all outputs are float32 numpy arrays, so convert types as appropriate
                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]

                output_dict_array.append(output_dict)
                dict_time.append(end - start)
    return output_dict_array, dict_time