Graph matching/pattern matching

[Tarar0011]

I am on working on the project human Gait recognition by accelerometer of smart phone. I successfully send data of accelerometer into the labview which i can see in the form of graph x,y,z with respect to time. Now the second step is to save that graph as reference wave and can be utilize for the authentication after matching. I am very new in Labview and i don't know how to do it . Please can any body send me tutorials or guide me how to do it. I shall be very thankful to you.

[VictorD]

Hello Tarar0011,

I am trying to understand what you are doing. So if you want to save the graph as a reference to have it as an image you should use an invoke node (to do it programatically) or to you can do it also manually.

This tutorial describes more about what I am talking about:  https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000YHq2CAG&l=en-US

Now, this link helps you to customize your chart/graph https://www.ni.com/docs/en-US/bundle/labview/page/customizing-graphs-and-charts.html

Further, you could save the data in an excel spreadsheet in case it is needed (Right-click over the graph/chart--Export--Export Data to Excel).

The project sounds really good. Actually I have studied the gait analysis for a short time and your project seems interesting. If you can provide me more info about the project, I would appreciate it just to learn about it. BTW are you using data dashboard? 

Hope I answered your question

[Tarar0011]

Thanks for the reply

Actually Gait analysis means human way of walking,every person have its own unique way of walking so we can utilize this characteristics in human authentication or identification or you can say that biometric security system. For Gait analysis there are several different techniques but  I have used accelerometer of smartphone (Android Galaxy S3) , accelerometer is a very sensitive sensor which measures magnitude and direction of the proper acceleration (or g-force), as a vector quantity, and can be used to sense orientation (because direction of weight changes), coordinate acceleration (so long as it produces g-force or a change in g-force), vibration and result can be shown on graph in x,y,z coordinates with respect to time. In short we can find three continuous waveforms of walking person, as it will be unique because every person have its unique Gait. So this will help in Identification of person

The question how data is transferred? Firstly I have android application which will send accelerometer data via Bluetooth to the computer then I have made VI which receives that data in real time and it shows on graphs. I hope you got my explanation if you need any more assistance please let me know. I am going to attach VI so you can understand(the VI recives data of others sensors as well so dont be confused).

Now my question is how can we utilize the imported graph (either in the form of image or in excel sheet) as a reference graph for identification/authentication or you we can say perform pattern matching?

[FuManWho]

Hey Tarar0011,

I found a similar question asked previously on this forum post about comparing waveform graphs.

Hopefully this will give you a good idea on resources to implement your application.

[Sara_snail]

I am working on gait analysis part for robotics, can you please share your full code with all the vi's it will be good for my reference. My email id: saravanaprashanth@outlook.com

[Bob_Schor]

If you really want to do analysis of your gait data, I would recommend saving the data, not a graph of the data.  I'm assuming that your data consist of a sampled set of X, Y, Z points, either saved as a 2D array of samples or as a 1D array of 3 Waveforms.  Note that either format can be easily written to disk (in various formats, depending on how you collect the data and how you want to store it), easily read back, and once in memory,  easily visualized on a Chart or Graph (not unlike what you did when you acquired the data).  By saving the data itself, you get the "purest", least-processed view of the data, ready for whatever comes next.

Bob Schor