danstowell / gmphd Goto Github PK
View Code? Open in Web Editor NEWGM-PHD filter implementation in python (Gaussian mixture probability hypothesis density filter)
License: GNU General Public License v3.0
GM-PHD filter implementation in python (Gaussian mixture probability hypothesis density filter)
License: GNU General Public License v3.0
====================================== gmphd GM-PHD filter implementation in python by Dan Stowell ====================================== This is a Python implementation of the Gaussian mixture PHD filter (probability hypothesis density filter) described in: B. N. Vo and W. K. Ma. The gaussian mixture probability hypothesis density filter. IEEE Transactions on Signal Processing, 54(11):4091--4104, 2006. DOI: 10.1109/TSP.2006.881190 It requires Numpy, and the demo scripts require matplotlib. Tested with Python 2.7. This implementation was developed as part of the following research: D. Stowell and M. D. Plumbley, Multi-target pitch tracking of vibrato sources in noise using the GM-PHD filter. In: Proceedings of Proceedings of the 5th International Workshop on Machine Learning and Music (MML12), July 2012. http://c4dm.eecs.qmul.ac.uk/papers/2012/StowellPlumbley2012mml.pdf The figures in that paper were produced by running the following commands: * Fig 1: `python syntheticexample.py` * Fig 2: `python syntheticroc.py` DIFFERENCES FROM VO & MA ======================== There are some differences from the GM-PHD algorithm described in Vo & Ma's paper: * I have not implemented "spawning" of new targets from old ones, since I don't need it. It would be straightforward to add it - see the original paper. * Weights are adjusted at the end of pruning, so that pruning doesn't affect the total weight allocation. * I provide an alternative approach to state-extraction (an alternative to Table 3 in the original paper) which makes use of the integral to decide how many states to extract. USAGE ===== The file "syntheticexample.py" is a python script which runs the filter over a synthetic randomly-generated scene, in which objects have 3D state and generate chirp-like observations. I suggest you start by looking at that script. But for a quick look at the API here's a very simple bit of python: from gmphd import * g = Gmphd([GmphdComponent(1, [100], [[10]])], 0.9, 0.9, [[1]], [[1]], [[1]], [[1]], 0.000002) g.update([[30], [67.5]]) g.gmmplot1d() g.prune() g.gmmplot1d() LICENCE ======= (c) 2012 Dan Stowell and Queen Mary University of London. All rights reserved. gmphd is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. gmphd is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with gmphd. If not, see <http://www.gnu.org/licenses/>.
The example seems to be 1d.
Hi @danstowell,
first of all thank you for making this project open-source I have been testing different implementations of GM-PHD filters. Yours also looks very promising. I was wondering if you took care of the data association problem in this implementation. Do you keep track of target IDs in some way? I would appreciate your help.
Cheers nwesem
Your code did not use pkk
though you calculated.
pkk = [dot(eye(len(k[index])) - dot(k[index], self.h), comp.cov)
object state [x y r h dx dy dr dh]
x, y are location of an object
r is object ration = width / height
h is object height
dx, dy, dr, dh are velocity of object w.r.t x,y,r,h
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ 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. ๐๐๐
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 โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.