Limited-Memory CMA-ES (LM-CMA-ES) is a variation of CMA-ES which has been tailored for large-scale optimisation with limited memory. As a warm up here, I have used this algorithm to solve the min point of the following function: $$ z= -( 4 \times e^{- ( (x-4)^2 +(y-4)^2 ) }+ 2 \times e^{- ( (x-2)^2 +(y-2)^2 ) } )$$ The function […]
Solving of optimisation problem with evolutionary algorithm (LM-CMA-ES) Read More »
This Matlab tutorial I use SIFT, RANSAC, and homography to find corresponding points between two images. Here I have used vlfeat to find SIFT features. Full code is available at my GitHub repository Major steps are: 0.Adding vlfeat to your Matlab workspace:
|
1 |
run('<path_to_vlfeat>/toolbox/vl_setup') |
1.Detect key points and extract descriptors. In the image below you can see some SIFT key
Stitching image using SIFT and Homography Read More »
Dynamic Time Warping (DTW) is a method to align two sequences such that they have minimum distance. For instance, two trajectories that are very similar but one of them performed in a longer time. Here is an example of my code with python. Here is my ROS package with C++ for DTW.
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 |
import numpy as np import matplotlib.pyplot as plt from scipy.spatial.distance import euclidean from fastdtw import fastdtw start=0 end=2*np.pi step=0.1 k=2 x1=np.arange(start,end,k*step) x2=np.arange(start,end/k,step) noise=np.random.uniform(start,end,len(x2))/10 y1=np.sin(x1)+1*np.sin(2*x1)+noise y2=np.sin(k*x2)+1*np.sin(k*2*x2) sin1=plt.plot(x1,y1) plt.setp(sin1,color='b',linewidth=2.0) sin2=plt.plot(x2,y2) plt.setp(sin2,color='r',linewidth=2.0) time_series_A=zip(x1,y1) time_series_B=zip(x2,y2) distance, path = fastdtw(time_series_A, time_series_B, dist=euclidean) print distance print path index_a,index_b=zip(*path) for i in index_a: x1=time_series_A[i][0] y1=time_series_A[i][1] x2=time_series_B[i][0] y2=time_series_B[i][1] plt.plot([x1, x2], [y1, y2], color='k', linestyle='-', linewidth=2) plt.show() |
Dynamic Time Warping with Python Read More »
In many applications, you need to compare two or more data sets with each other to see how much they are similar or different. For instance, you have measured the height of men and women in Japan and the Netherlands and now you like to know how much they are different. Two commonly used method
Analytic distance metric for Gaussian mixture models Read More »
In this work at first, I recognize the object in the scene and estimate the 6 DOF pose of that. Then I track the object by using particle filter. RGB data acquired from Kinect 2 and turned into PCL pointcloud. I demonstrate a task several times to the robot. In this case, I move an object (a detergent) over
Learning From Demonstration Read More »
