imsoo / fight_detection Goto Github PK
View Code? Open in Web Editor NEWReal time Fight Detection Based on 2D Pose Estimation and RNN Action Recognition
License: MIT License
Real time Fight Detection Based on 2D Pose Estimation and RNN Action Recognition
License: MIT License
fps
Hi, I found your work very interesting and I try to run a training by myself.
However, I didn't find your test samples. Could you please upload it to Google Drive just like the training samples. Thanks!
Hi i download the feature vector dataset from the link:https://drive.google.com/file/d/1ZNJDzQUjo2lDPwGoVkRLg77eA57dKUqx/view
the content of the download feature vector like this:
182.121,0.0372238,94.563,316.817,0.0580139,88.1386,76.9897,0.0069809,
85.9977,33.9192,0.0205917,80.3989,0,0,0,0,
0,0,0,0,0,0,0,0
182.137,0.540245,94.0227,316.774,0.898399,89.037,76.9441,0.00663757,
86.0043,36.2493,0.269905,80.6688,0,0,0,0,
0,0,0,0,0,0,0,0
i think the second line of the second feature vector must not the Δangle, so i check the all of dataset.i guess i nedd a description of the dataset feature vector format. thanks
Is this project working? If so, this is exciting! I'm going to give it a try
tensorflow.python.framework.errors_impl.InvalidArgumentError: No OpKernel was registered to support Op 'CudnnRNN' used by node cu_dnnlstm/CudnnRNN (defined at action.py:28) with these attrs: [seed=0, dropout=0, input_mode="linear_input", T=DT_FLOAT, direction="unidirectional", rnn_mode="lstm", seed2=0, is_training=true]
Registered devices: [CPU]
Registered kernels:
[[cu_dnnlstm/CudnnRNN]]
Hi @imsoo, please let me know how do you train your lstm, a specific command would be helpful. I intend to prepare my own datasets with custom actions.
What's the unit of Δpoint? Is it presented as number of pixels? If so, may I ask why it is not normalized by the width and height of the video? I believe it will be more robust if it is irrelevant to the size of the video.
And I assume that the units of angle and Δangle are degrees but not rads. Is that correct?
i build the server success, but when i run all pipeline, i get nothing display on the output window except the "OUTPU". something wrong? i read the code, i find the openpose works correctly, but i can't find the action.
Hi, imsoo.
First, thank you for providing me with a good repo.
As the title says, I want this repo to work as a local file.
So I didn't operate ventilator, worker, and client.
If so, I think I can only use action.py and OpenCV, but I would like to know if it is possible.
My current development environment is Python 3.6 and tf 2.0, and I am going to load the pre trained action.h5 model and apply it to the image.
Thank you for you read my question.
Hi, I've got these data (171204_pose3, 171204_pose5, 171204_pose6) from the CMU Panoptic Dataset.
However, the punch data was not found. It seems that Berkeley mhad data can only be downloaded completely, which is too big. Therefore, can you provide the video data of punch. thank you
Hi, I found some questions about the order of your data.
I guess they are organized by the order of (delta dist, delta angle, angle) for each of the 8 items? And I also found it's quite difficult to understand your 'delta dist', what's your measurement unit?I found my 'delta dist' very small. And is it really the distance of one point between current frame and previous frame? I think it too big, and I think it is the distance between different points in the same frame.
Thank you.
When i run the command,
./worker cfg/openpose.cfg weights/openpose.weights -gpu 0 -pose
I am getting this error "libdarknet.so: cannot open shared object file: No such file or directory". But I have compiled darknet in the way mentioned in the repo. Looks like fight detection Makefile is not updated with correct libraries?
I am very interested in this project. Could you please check the error msg? I build server part successfully, but stuck with the client.
/usr/include/x86_64-linux-gnu/sys/select.h:78:0: note: this is the location of the previous definition
#define FD_SETSIZE __FD_SETSIZE
^
./src/main.cpp:87:10: error: aggregate ‘std::ofstream out_json_file’ has incomplete type and cannot be defined
ofstream out_json_file;
^
./src/main.cpp: In function ‘int main(int, char**)’:
./src/main.cpp:142:13: error: ‘CAP_PROP_FPS’ was not declared in this scope
cap.set(CAP_PROP_FPS, 20);
^
./src/main.cpp:143:13: error: ‘CAP_PROP_BUFFERSIZE’ was not declared in this scope
cap.set(CAP_PROP_BUFFERSIZE, 3);
^
./src/main.cpp:159:24: error: ‘CAP_PROP_FPS’ was not declared in this scope
double fps = cap.get(CAP_PROP_FPS);
^
./src/main.cpp:160:23: error: ‘CAP_PROP_FRAME_COUNT’ was not declared in this scope
end_frame = cap.get(CAP_PROP_FRAME_COUNT);
^
./src/main.cpp:176:43: error: invalid use of incomplete type ‘std::ofstream {aka class std::basic_ofstream}’
out_json_file = ofstream(out_json_path);
^
In file included from /usr/include/c++/5/ios:38:0,
from /usr/include/c++/5/ostream:38,
from /usr/include/c++/5/iostream:39,
from ./src/main.cpp:2:
/usr/include/c++/5/iosfwd:119:11: note: declaration of ‘std::ofstream {aka class std::basic_ofstream}’
class basic_ofstream;
^
./src/main.cpp:185:31: error: ‘fourcc’ is not a member of ‘cv::VideoWriter’
writer.open(out_vid_path, VideoWriter::fourcc('M', 'P', '4', 'V'), fps, Size(cap_width, cap_height), true);
^
Makefile:35: recipe for target 'obj/main.o' failed
make: *** [obj/main.o] Error 1
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.