DTW computation
Project description
##Dynamic Time Warping Project
[Examples](http://nbviewer.ipython.org/github/honeyext/cdtw/blob/master/examples.ipynb)
This module implements:
Distance functions:
* manhattan
* euclidean
* squared euclidean
Local constraints(step patterns, step functions):
* [well known step patterns dp1, dp2, dp3][1]
* [local constraints classified by Sakoe-Chiba][2]
impo
Global constraints(windows):
* Itakura parallelogram
* [Sakoe-chiba band, Palival adjustment window][3]
```python
import numpy as np
from cdtw import pydtw
r = np.array([1,2,3,4])
q = np.array([2,3,4,5])
d = pydtw.dtw(r,q,pydtw.Settings(step = 'p0sym', #Sakoe-Chiba symmetric step with slope constraint p = 0
window = 'palival', #type of the window
param = 2.0, #window parameter
norm = False, #normalization
compute_path = True))
d.get_dist()
#2.0
d.get_cost()
#array([[ 1., 3., 6., inf],
# [ 1., 2., 4., 7.],
# [ 2., 1., 2., 4.],
# [ inf, 2., 1., 2.]])
d.get_path()
#[(0, 0), (1, 0), (2, 1), (3, 2), (3, 3)]
```
[1]: http://cyber.felk.cvut.cz/gerstner/teaching/zbd/dtw.pdf
[2]: http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=1163055&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D1163055
[3]: https://maxwell.ict.griffith.edu.au/spl/publications/papers/sigpro82_kkp_dtw.pdf
[Examples](http://nbviewer.ipython.org/github/honeyext/cdtw/blob/master/examples.ipynb)
This module implements:
Distance functions:
* manhattan
* euclidean
* squared euclidean
Local constraints(step patterns, step functions):
* [well known step patterns dp1, dp2, dp3][1]
* [local constraints classified by Sakoe-Chiba][2]
impo
Global constraints(windows):
* Itakura parallelogram
* [Sakoe-chiba band, Palival adjustment window][3]
```python
import numpy as np
from cdtw import pydtw
r = np.array([1,2,3,4])
q = np.array([2,3,4,5])
d = pydtw.dtw(r,q,pydtw.Settings(step = 'p0sym', #Sakoe-Chiba symmetric step with slope constraint p = 0
window = 'palival', #type of the window
param = 2.0, #window parameter
norm = False, #normalization
compute_path = True))
d.get_dist()
#2.0
d.get_cost()
#array([[ 1., 3., 6., inf],
# [ 1., 2., 4., 7.],
# [ 2., 1., 2., 4.],
# [ inf, 2., 1., 2.]])
d.get_path()
#[(0, 0), (1, 0), (2, 1), (3, 2), (3, 3)]
```
[1]: http://cyber.felk.cvut.cz/gerstner/teaching/zbd/dtw.pdf
[2]: http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=1163055&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D1163055
[3]: https://maxwell.ict.griffith.edu.au/spl/publications/papers/sigpro82_kkp_dtw.pdf
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
cdtw-0.0.1.zip
(181.7 kB
view hashes)