VOC转tfrecord

ubuntu16.04 系统

使用Tensorflow Object Detection API 将VOC转化到tfrecord

1. 准备

pip install tensorflow # 1.7

2. 下载 tensoflow下的model模块

https://github.com/tensorflow/models

3. 安装models中需要的一些环境

参考 https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md

3.1 配置probuf

# From tensorflow/models/research/
wget -O protobuf.zip https://github.com/google/protobuf/releases/download/v3.0.0/protoc-3.0.0-linux-x86_64.zip
unzip protobuf.zip
./bin/protoc object_detection/protos/*.proto --python_out=.

3.2 Add Libraries to PYTHONPATH

# From tensorflow/models/research/
export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim

4. 下载VOC2007数据集

http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar

解压后放在 /home 路径下

5. 修改一下creat_pascal_tf_record.py 中的相关路径

# flags.DEFINE_string('label_map_path', 'data/pascal_label_map.pbtxt',
#                     'Path to label map proto')
flags.DEFINE_string('label_map_path', '/home/xx/cv_xx/models/research/object_detection/data/pascal_label_map.pbtxt',
                    'Path to label map proto')

6. 开始转换

      python object_detection/dataset_tools/create_pascal_tf_record.py \
        --data_dir=/home/xx/VOCdevkit \
        --year=VOC2007 \
        --output_path=/home/xx/pascal.record