Tensorflow Object Detection API介绍和应用(mac版)——（3.加载预训练模型训练自定义模型）

1、前言

本篇用到的pre-trained model是coco（上一篇提到的），下载几个图片文件作为训练数据，迭代次数10次，训练差不都5分钟。只是给个例子，本机小破mac怕爆炸。看google建议是每个label是100张图片（能500+最好）。本人实测，一个label样式不是很多的logo类图像，40或50个训练数据，200k迭代，感觉效果已经不错了。
上一篇：https://blog.youkuaiyun.com/bjjoy2009/article/details/94962333

2、安装cocoapi

打开终端，按顺序执行下面操作
（1）我的本地cocoapi下载路经
cd /Users/bjjoy
（2）git clone https://github.com/cocodataset/cocoapi.git
（3）cd cocoapi/PythonAPI
（4）make
（5）将编译好的文件夹放到第一篇下载的models对应路径下
cp -r pycocotools /Users/bjjoy/JupyterProjects/models/research/

3、安装LabelImg

该项目就是标记图像，打标签用的(python3以上)，直接终端安装
（1）pip install labelImg
（2）终端输入：labelImg，打开下图

4、构建测试数据

（1）在model路径下（~/JupyterProjects/models/research/object_detection）创建3个文件夹

my_train：存放训练模型
my_test_data: 训练和测试图像数据
my_image：需要标注的图像数据（这里用了6个huawei的logo，4训练2测试）
（2）用labelImg标记图像（这里标签就用的huawei）

左侧绿框 Open Dir：下载的图片存放路径（这里是my_images）
左侧红框Create\nRecBox：标记图像（右侧绿色框）
输入标签名：点击ok
保存（command+s)：存储为xml文件，每个图像一个xml文件

5、训练数据xml转为csv

参考：https://github.com/datitran/raccoon_dataset
（1）xml ->csv
用pycharm创建了一个xml转csv的项目，结构如下图

将labelImg生成的xml文件放到绿色框对应文件夹中
红色框xml_to_csv.py程序如下，运行后csv文件放到data目录下

import os
import glob
import pandas as pd
import xml.etree.ElementTree as ET

def xml_to_csv(path):
    xml_list = []
    for xml_file in glob.glob(path + '/*.xml'):
        tree = ET.parse(xml_file)
        root = tree.getroot()
        for member in root.findall('object'):
            value = (root.find('filename').text,
                     int(root.find('size')[0].text),
                     int(root.find('size')[1].text),
                     member[0].text,
                     int(member[4][0].text),
                     int(member[4][1].text),
                     int(member[4][2].text),
                     int(member[4][3].text)
                     )
            xml_list.append(value)
    column_name = ['filename', 'width', 'height', 'class', 'xmin', 'ymin', 'xmax', 'ymax']
    xml_df = pd.DataFrame(xml_list, columns=column_name)
    return xml_df

def main2():
    for directory in ['train', 'test']:
        image_path = os.path.join(os.getcwd(), '{}'.format(directory))
        xml_df = xml_to_csv(image_path)
        xml_df.to_csv('data/{}_labels.csv'.format(directory), index=None)
        print('Successfully converted xml to csv.')

main2()

6、csv转为tfrecord（训练识别的文件类型）

（1）将上面的csv文件搬到models/research/object_detection/my_test_data
（2）在models/research/object_detection 创建一个generate_tfrecord.py，代码如下

from __future__ import division
from __future__ import print_function
from __future__ import absolute_import

import os
import io
import pandas as pd
import tensorflow as tf

from PIL import Image
from object_detection.utils import dataset_util
from collections import namedtuple, OrderedDict

flags = tf.app.flags
flags.DEFINE_string('csv_input', '', 'Path to the CSV input')
flags.DEFINE_string('output_path', '', 'Path to output TFRecord')
flags.DEFINE_string('image_dir', '', 'Path to images')
FLAGS = flags.FLAGS


# TO-DO replace this with label map
def class_text_to_int(row_label):
    if row_label == 'huawei':
        return 1
    else:
        None


def split(df, group):
    data = namedtuple('data', ['filename', 'object'])
    gb = df.groupby(group)
    return [data(filename, gb.get_group(x)) for filename, x in zip(gb.groups.keys(), gb.groups)]


def create_tf_example(group, path):
    with tf.gfile.GFile(os.path.join(path, '{}'.format(group.filename)), 'rb') as fid:
        encoded_jpg = fid.read()
    encoded_jpg_io = io.BytesIO(encoded_jpg)
    image = Image.open(encoded_jpg_io)
    width, height = image.size

    filename = group.filename.encode('utf8')
    image_format = b'jpg'
    xmins = []
    xmaxs = []
    ymins = []
    ymaxs = []
    classes_text = []
    classes = []

    for index, row in group.object.iterrows():
        xmins.append(row['xmin'] / width)
        xmaxs.append(row['xmax'] / width)
        ymins.append(row['ymin'] / height)
        ymaxs.append(row['ymax'] / height)
        classes_text.append(row['class'].encode('utf8'))
        classes.append(class_text_to_int(row['class']))

    tf_example = tf.train.Example(features=tf.train.Features(feature={
        'image/height': dataset_util.int64_feature(height),
        'image/width': dataset_util.int64_feature(width),
        'image/filename': dataset_util.bytes_feature(filename),
        'image/source_id': dataset_util.bytes_feature(filename),
        'image/encoded': dataset_util.bytes_feature(encoded_jpg),
        'image/format': dataset_util.bytes_feature(image_format),
        'image/object/bbox/xmin': dataset_util.float_list_feature(xmins),
        'image/object/bbox/xmax': dataset_util.float_list_feature(xmaxs),
        'image/object/bbox/ymin': dataset_util.float_list_feature(ymins),
        'image/object/bbox/ymax': dataset_util.float_list_feature(ymaxs),
        'image/object/class/text': dataset_util.bytes_list_feature(classes_text),
        'image/object/class/label': dataset_util.int64_list_feature(classes),
    }))
    return tf_example


def main(_):
    writer = tf.python_io.TFRecordWriter(FLAGS.output_path)
    path = os.path.join(FLAGS.image_dir)
    examples = pd.read_csv(FLAGS.csv_input)
    grouped = split(examples, 'filename')
    for group in grouped:
        tf_example = create_tf_example(group, path)
        writer.write(tf_example.SerializeToString())

    writer.close()
    output_path = os.path.join(os.getcwd(), FLAGS.output_path)
    print('Successfully created the TFRecords: {}'.format(output_path))


if __name__ == '__main__':
    tf.app.run()

（3）生成tfrecord
i）打开终端，cd进入到models/research
ii) 切换python项目路径：export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim
iii)cd进入models/research/object_detection，执行下面2条命令：

python generate_tfrecord.py --csv_input=my_test_data/test_labels.csv --output_path=my_test_data/test.record --image_dir=my_images/

python generate_tfrecord.py --csv_input=my_test_data/train_labels.csv --output_path=my_test_data/train.record --image_dir=my_images/

7、训练

7.1 生成标签文件

在my_test_data文件夹创建huawei_detetion.pbtxt，内容如下

item {
id: 1
name: ‘huawei’
}

7.2 解压coco模型内文件

此处用的版本是ssd_mobilenet_v1_coco_2017_11_17.tar.gz，直接解压到object_detection文件夹下即可。

7.3 配置config

（1）在models/research/object_detection/samples/configs，找到ssd_mobilenet_v1_coco.config，复制到my_test_data中
（2）修改config文件，主要修改如下
迭代次数，num_steps: 10
减小 batch_size: 24
文件最下面，所有路径相关参数，例如 fine_tune_checkpoint

# SSD with Mobilenet v1 configuration for MSCOCO Dataset.
# Users should configure the fine_tune_checkpoint field in the train config as
# well as the label_map_path and input_path fields in the train_input_reader and
# eval_input_reader. Search for "PATH_TO_BE_CONFIGURED" to find the fields that
# should be configured.

model {
  ssd {
    num_classes: 1
    box_coder {
      faster_rcnn_box_coder {
        y_scale: 10.0
        x_scale: 10.0
        height_scale: 5.0
        width_scale: 5.0
      }
    }
    matcher {
      argmax_matcher {
        matched_threshold: 0.5
        unmatched_threshold: 0.5
        ignore_thresholds: false
        negatives_lower_than_unmatched: true
        force_match_for_each_row: true
      }
    }
    similarity_calculator {
      iou_similarity {
      }
    }
    anchor_generator {
      ssd_anchor_generator {
        num_layers: 6
        min_scale: 0.2
        max_scale: 0.95
        aspect_ratios: 1.0
        aspect_ratios: 2.0
        aspect_ratios: 0.5
        aspect_ratios: 3.0
        aspect_ratios: 0.3333
      }
    }
    image_resizer {
      fixed_shape_resizer {
        height: 300
        width: 300
      }
    }
    box_predictor {
      convolutional_box_predictor {
        min_depth: 0
        max_depth: 0
        num_layers_before_predictor: 0
        use_dropout: false
        dropout_keep_probability: 0.8
        kernel_size: 1
        box_code_size: 4
        apply_sigmoid_to_scores: false
        conv_hyperparams {
          activation: RELU_6,
          regularizer {
            l2_regularizer {
              weight: 0.00004
            }
          }
          initializer {
            truncated_normal_initializer {
              stddev: 0.03
              mean: 0.0
            }
          }
          batch_norm {
            train: true,
            scale: true,
            center: true,
            decay: 0.9997,
            epsilon: 0.001,
          }
        }
      }
    }
    feature_extractor {
      type: 'ssd_mobilenet_v1'
      min_depth: 16
      depth_multiplier: 1.0
      conv_hyperparams {
        activation: RELU_6,
        regularizer {
          l2_regularizer {
            weight: 0.00004
          }
        }
        initializer {
          truncated_normal_initializer {
            stddev: 0.03
            mean: 0.0
          }
        }
        batch_norm {
          train: true,
          scale: true,
          center: true,
          decay: 0.9997,
          epsilon: 0.001,
        }
      }
    }
    loss {
      classification_loss {
        weighted_sigmoid {
        }
      }
      localization_loss {
        weighted_smooth_l1 {
        }
      }
      hard_example_miner {
        num_hard_examples: 3000
        iou_threshold: 0.99
        loss_type: CLASSIFICATION
        max_negatives_per_positive: 3
        min_negatives_per_image: 0
      }
      classification_weight: 1.0
      localization_weight: 1.0
    }
    normalize_loss_by_num_matches: true
    post_processing {
      batch_non_max_suppression {
        score_threshold: 1e-8
        iou_threshold: 0.6
        max_detections_per_class: 100
        max_total_detections: 100
      }
      score_converter: SIGMOID
    }
  }
}

train_config: {
  batch_size: 24
  optimizer {
    rms_prop_optimizer: {
      learning_rate: {
        exponential_decay_learning_rate {
          initial_learning_rate: 0.004
          decay_steps: 800720
          decay_factor: 0.95
        }
      }
      momentum_optimizer_value: 0.9
      decay: 0.9
      epsilon: 1.0
    }
  }
  fine_tune_checkpoint: "object_detection/ssd_mobilenet_v1_coco_2017_11_17/model.ckpt"
  from_detection_checkpoint: true
  # Note: The below line limits the training process to 200K steps, which we
  # empirically found to be sufficient enough to train the pets dataset. This
  # effectively bypasses the learning rate schedule (the learning rate will
  # never decay). Remove the below line to train indefinitely.
  num_steps: 10
  data_augmentation_options {
    random_horizontal_flip {
    }
  }
  data_augmentation_options {
    ssd_random_crop {
    }
  }
}

train_input_reader: {
  tf_record_input_reader {
    input_path: "object_detection/my_test_data/train.record"
  }
  label_map_path: "object_detection/my_test_data/huawei_detetion.pbtxt"
}

eval_config: {
  num_examples: 12
  # Note: The below line limits the evaluation process to 10 evaluations.
  # Remove the below line to evaluate indefinitely.
  max_evals: 10
}

eval_input_reader: {
  tf_record_input_reader {
    input_path: "object_detection/my_test_data/test.record"
  }
  label_map_path: "object_detection/my_test_data/huawei_detetion.pbtxt"
  shuffle: false
  num_readers: 1
}

8、训练

终端执行，等几分钟
python object_detection/model_main.py --model_dir=object_detection/my_train/ --pipeline_config_path=object_detection/my_test_data/ssd_mobilenet_v1_coco.config

9、tensorboard查看结果

tensorboard --logdir=‘object_detection/my_train’

10、object_detection_tutorial.ipynb测试训练好的模型

（1）模型转为pb文件
python object_detection/export_inference_graph.py --input_type image_tensor --pipeline_config_path object_detection/my_test_data/ssd_mobilenet_v1_coco.config --trained_checkpoint_prefix object_detection/my_train/model.ckpt-10 --output_directory object_detection/my_test_data/frozen_inference_graph
（2）object_detection_tutorial.ipynb运行
修改模型路径：

PATH_TO_FROZEN_GRAPH = 'my_test_data' + '/frozen_inference_graph.pb'
PATH_TO_LABELS = os.path.join('my_test_data', 'huawei_detetion.pbtxt')

Download Model模块代码全部注释掉

# opener = urllib.request.URLopener()
# opener.retrieve(DOWNLOAD_BASE + MODEL_FILE, MODEL_FILE)
# tar_file = tarfile.open(MODEL_FILE)
# for file in tar_file.getmembers():
#   file_name = os.path.basename(file.name)
#   if 'frozen_inference_graph.pb' in file_name:
#     tar_file.extract(file, os.getcwd())

测试图片参数

TEST_IMAGE_PATHS = ['my_images/hw8.jpg']

（3）运行结果，及其不准

11、总结

写的有点长，中间也不知道哪里会有问题了，有些是根据回忆。只是没想到原来想用现成的项目原来也这么麻烦啊。