I.GitHub project
main project: https://github.com/danfeiX/scene-graph-TF-release
dataset prepare:https://github.com/danfeiX/scene-graph-TF-release/tree/master/data_tools
II.配置记录 configuration records:
python版本：2.7
tensorflow版本：0.11.0
cuda版本：8.0
cudnn版本：7.0.5
III.Visual Genmoe数据集与scene-graph项目
1.VGdataset数据格式：
https://blog.csdn.net/zziahgf/article/details/79097839
注：object_id, relationship_id等为某一类的id，并非单个物体，例如所有"man"的id均为1058532，而不是每个人不同。
objects.json的数据格式的例子，换行表示
[
{"image_id": 1,
"objects":
[{"synsets": ["tree.n.01"],
"h": 557,
"object_id": 1058549,
"merged_object_ids": [],
"names": ["trees"]
"w": 799,
"x":0,
"y":0},...]
...}
{"image_id:" 2...
}...], "image_url": ""}
]
注，boudingbox的x、y应指其左上角坐标
2.项目所需数据：
image，格式jpg
image metadata
vg scene graph其中包括object和relationships，格式均为json
3.relationships.json格式
下面为VG数据集中的一个图片，其image_id为1
所有图片的relationships的集合，即relationships.json的格式(例子)为：
[
{"relationships":
[{relationship描述1}, {relationship描述2}...],
"image_id": 1}, ...
]
而 relationship描述 的格式(例子)为：
{
"predicate": "wears",
"object": {
"name": "sneakers",
"h": 26,
"object_id": 5048,
"synsets": [gym_shoe.n.01"],
"w": 52,
"y": 489,
"x": 243 },
"relationshp_id": 15937,
"synsets": ["wear.v.01"],
"subject": {
"name": "man",
"h": 262,
"object_id": 1058529,
"synsets": ["man.n.01"],
"w": 60,
"y": 249,
"x": 238 },
}
3.sh文件所需文件：
create_imdb.sh：输入：jpg图片及其metadata; 输出：图片h5格式的文件，imdb_1024.h5
create_rodib.sh：输入