gjy3035 / gcc-cl Goto Github PK
View Code? Open in Web Editor NEWGCC dataset Collector and Labeler (GCC CL) [CVPR2019]
Home Page: https://gjy3035.github.io/GCC-CL/
License: MIT License
GCC dataset Collector and Labeler (GCC CL) [CVPR2019]
Home Page: https://gjy3035.github.io/GCC-CL/
License: MIT License
因为nuget安装的minhook库无法使用,我用了GCC-CL-TOOLS里面的GCC-CL.asi,设置相机,多边形区域和四个相机还有设置人数范围都是正常的,但是重开游戏按L后,在生成人物时会显示GCC-CL.asi出错,请问是因为不能直接用GCC-CL.asi吗,另外,我没有安装No Chromatic aberration & Lens distortion 1.52,会对运行有影响吗
I see relavant functions in export.cpp? But how to use them?
Hi, I have created and saved 1 scene by following your steps. I restart the game and press L
, then my man falls from the sky to ground successfully. But then(like 3:44 in your youtube video) there comes an error window:
SCRIPT HOOK V ERROR
CORE: An exception occurred while executing 'GCC-CL.asi'
(0x00007FFC0ED7D890), id43
, and there are no people on the ground. I can do nothing but exit the game.
Could you tell me what's happening? If I want to debug it, which source file should I look at?
When I complie GCC-Collector.sln, it shows like that "ERROR LNK1181: unable to open input file “...\GCC_CL\GCC-CL\GCC-CL-master\GCC-Collector\packages\minhook.1.3.3\build\native....\lib\native\lib\libMinHook.lib” "
How can I solve this problem?
Can you give me GCC-Collector.asi?
Hi, I have successfully generated scenes. I want to generate binary masks of people then. So, I directly run the main.py
in GCC-Labeler, then read and show the .raw
file:
raw_path = 'D:/Codes/gta5dataset/scene_1_0/segs/1558631488.raw'
seg = np.fromfile(raw_path, 'uint8')
seg = seg.reshape(1440, 2560) # I use a 2k resollution GTA V
plt.imshow(s,cmap='gray')
plt.show()
Then the figure is like this:
The original GTA V scene is:
It seems that there is something wrong. I also visualized the GTA V/data/part_1_0/xxxxxxxx/part_0.raw
(.raw file created by GCC-Collector), and it is the same as the above black&white image.
Could you tell me where should I look at to solve this problem?
Hello author, why did I press F9 not respond. The configuration process has been strictly followed.
hello! I have a question regarding the format of the JSON files
"camera": {"location": [-1429.03, -969.06, 32.7857], "rotation": [-19.942, 1.10123e-05, 92.217], "fov": 40.0}}
for the camera location, does it represent X, Y and Z value?
Is the Z value the height in terms of metres?
For the FOV value, is that the pitch angle down from the horizontal view?
Or is the rotation value, which of them is the pitch angle?
Thanks!
I created data file right and part_0_1 to part_0_3 occurs,
But when I press 'L' , error occurs from GTA5 like this:
CORE: An exception occurred while executing 'GCC-Collector.asi'
(0x00007FFC9F38D0E0), id 33
and then GTA5 crashed.
GCC-CL.log is:
[1667660082468] : get eye info done.
[1667660082468] : -21.427000 -1468.140015 33.252602
[1667660082468] : -18.879000 -0.000008 93.644997
[1667660082468] : 40.000000
[1667660082568] : finished keep camera...
[1667660091622] : end defineArea
[1667660106833] : now have = -1
I tried the methods provided in issue#24 and issue#5 but they didn't fix my error.
Can we generate videos with people masked?
In Script.cpp line 26, function "initAnimations " get "SceneDirectorAnim.txt". Unfortunately, I cannot find the txt file so that the function always returned False. Can you offer the file? Thanks!
Hi, in your project website you show a binary mask of people in addition to marks on people's head. How can we achieve that using GTA V?
This is a great work!
I have a doubt about using it. I cannot stop the collecting process, and the character keep jumping from the sun after every few snapshots.
I wonder how to stop the process in the middle, and if it is normal for the character to keep bungee-jumping?
Thanks~
very nice work!!! can you tell me the version of minhook and eigen you used? and which c++ standard you used, C++11 or C++ 14 Or C++17?
Thanks for your amazing work.
Recently, I am interested in this project, and I really want to set up this system to support my study.
But I encounter some issues:
Thanks a lot!
when I try to compile noVehicle,VS report error: please define entry point first.what happend?
Python scrpits error \GCC-CL\GCC-Labeler\combine.py", line 39, in backMask mask = np.where(diff > rgbDis, o, mask) ValueError: operands could not be broadcast together with shapes (1080,2048) (1080,1920)
And I also find out my collected image folder struct like :
/data/part_1_0/1567414646/
part_0.raw
part_0_0.bmp
part_0_1.bmp
pedinfo.xml
only 4 files.
Why missing files about 180 and 360? Is it because something wrong with camera?
Is the mod "No Chromatic aberration & Lens distortion" is necessary? I didn't install it.
I readly appreciate it if anyone could reply.
I have compiled GCC-Collector.asi, put all necessary files (such as SceneDirectorAnim.txt) in game’s root folder, and followed the instruction carefully (choosing level 3). When I restart and press ‘L’, the main character starts falling, camera moves to the selected scene and a set of characters is generated. But then the script freezes with no error. It just stops generating new characters and only 1 file is created (with no images or meta)
The exact structure of my data folder is below:
data part_1_0 1688582167 // for some reason such a folder was generated only for part 1_0 pedInfo.xml // and only 1 file has been generated here anyway, with no images\meta areaInfo.log eyeInfo.log levelInfo.log Zheight.log part_1_1 areaInfo.log eyeInfo.log levelInfo.log Zheight.log part_1_2 areaInfo.log eyeInfo.log levelInfo.log Zheight.log part_1_3 areaInfo.log eyeInfo.log levelInfo.log Zheight.log
I have also tried using the pre-compiled GCC-CL.asi from gcc tools, but always got error when pressing ‘L’ like in issues #24 and #5, so I had to recompile it.
Could it be a limitation of my machine, such as not enough memory for the game and the script to work properly?
I would really appreciate your help with repairing the script. My logs will be posted below
Hello! I found a possible issue with one of the images in the dataset. Please have a look at image 1532170935
it is a black image, i assume it is a loading screen because of the white circle on the bottom right
the ground truth given is 24, but there is nothing in the image since it is a black loading screen. Could rectify, please. Thank you!
Hi @gjy3035 thanks for your work. This is a really awesome work in synthetic dataset.
I want to ask, is there a way to produce thermal view of the crowd?
Something like that, where we can choose the colour palette. Preferably WhiteHot.
I understnad that it is likely to need a mod. Do you know where to find such a mod to get the Thermal camera views?
Once again, thank you
After executing Python script without any error, only some normal GTA screenshot generated, but where's the mask picture?
I got output folders as:
jpgs
jsons
mats
pngs
vis
But I can't get images with people masked. I can get combination.bmp
under gta data folder, I check source code, the result.jpg
will be renamed and moved to the VIS
folder, but still they're not mask image...
collected logs here data.zip
And I also find out my collected image folder struct like :
/data/part_1_0/1567414646/
part_0.raw
part_0_0.bmp
part_0_1.bmp
pedinfo.xml
only 4 files.
Why missing files about 180 and 360? Is it because something wrong with camera?
I extactly follow the instructions, except for installing the mod No Chromatic aberration & Lens distortion
, noVehicle
, unlimitedLife
I really appreciate it if anyeone could reply anything.
If the background or scene can be changed using self-image-data?
I can set camera and points but when I push 'L'. It crashes and gives the error:
CORE: An exception occurred while executing 'GCC-CL.asi....
Id is different for different trials.
Here is my log:
GCC-CL.log
End of the log:
[1640546796542] : finished keep camera...
[1640546812921] : data\part_1_3\eyeInfo.log
[1640546812922] : get eye info done.
[1640546812922] : -46.131100 461.612000 159.093002
[1640546812922] : -40.060200 0.000015 210.817993
[1640546812922] : 40.000000
[1640546813048] : finished keep camera...
[1640546822267] : end defineArea
[1640546850844] : now have = -1
I have read the issue #5. My txt file is in the root folder and some files(part_ folders and .logs extension files) are created in my data folder.
What could be the cause of the error? Could you help me?
错误 C1083 无法打开包括文件: “natives.h”: No such file or directory GCC-Collector c:\users\51604\desktop\gcc-cl-master\gcc-collector\gcc-collector\animation.cpp 6
错误 C1083 无法打开包括文件: “nativeCaller.h”: No such file or directory GCC-Collector c:\users\51604\desktop\gcc-cl-master\gcc-collector\gcc-collector\export.cpp 2
错误 C1083 无法打开包括文件: “main.h”: No such file or directory GCC-Collector c:\users\51604\desktop\gcc-cl-master\gcc-collector\gcc-collector\main.cpp 1
错误 C1083 无法打开包括文件: “types.h”: No such file or directory GCC-Collector c:\users\51604\desktop\gcc-cl-master\gcc-collector\gcc-collector\utils.h 12
……
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.