My educational project is about "sign language recognition using kinect camera" .
I want to compare the hand motion trajectories using DTW as the distance measure , and then do a NN-DTW classification .
Hand trajectory is constructed from the hand joint position in consecutive frames in 3D coordinate system.
x,y,z coordinate for hand joint , in every frame , is obtained from using kinect camera .
Which option is more appropriate for measuring the distance of these trajectories? DTWi or DTWd ?
Short answer: DTWd (for your specific use-case)
You might want to have a look at this paper:
Shokoohi-Yekta, M., Wang, J., & Keogh, E. (2015). On the Non-Trivial Generalization of Dynamic Time Warping to the Multi-Dimensional Case. Proceedings of the 2015 SIAM International Conference on Data Mining, 289–297. https://doi.org/10.1137/1.9781611974010.33
According to this paper, one of the two (i.e., DTWi or DTWd) always exhibits better performance. However, the decision will depend on the data you are using. In general terms, the authors say that "results suggest if the data dimensions are dependently warped, use DTWd to classify the data. If the data dimensions are independently warped, DTWi will give you more accurate results for classifying the data"
If you do not want to make that decision before hand, you can continuously adapt the selection to the one that suits best as explained in their paper.
Related
I'm trying to build a regression based M/L model using tensorflow.
I am trying to estimate an object's ETA based on the following:
distance from target
distance from target (X component)
distance from target (Y component)
speed
The object travels on specific journeys. This could be represented as from A->B or from A->C or from D->F (POINT 1 -> POINT 2). There are 500 specific journeys (between a set of points).
These journeys aren't completely straight lines, and every journey is different (ie. the shape of the route taken).
I have two ways of getting around this problem:
I can have 500 different models with 4 features and one label(the training ETA data).
I can have 1 model with 5 features and one label.
My dilemma is that if I use option 1, that's added complexity, but will be more accurate as every model will be specific to each journey.
If I use option 2, the model will be pretty simple, but I don't know if it would work properly. The new feature that I would add are originCode+ destinationCode. Unfortunately these are not quantifiable in order to make any numerical sense or pattern - they're just text that define the journey (journey A->B, and the feature would be 'AB').
Is there some way that I can use one model, and categorize the features so that one feature is just a 'grouping' feature (in order separate the training data with respect to the journey.
In ML, I believe that option 2 is generally the better option. We prefer general models rather than tailoring many models to specific tasks, as that gets dangerously close to hardcoding, which is what we're trying to get away from by using ML!
I think that, depending on the training data you have available, and the model size, a one-hot vector could be used to describe the starting/end points for the model. Eg, say we have 5 points (ABCDE), and we are going from position B to position C, this could be represented by the vector:
0100000100
as in, the first five values correspond to the origin spot whereas the second five are the destination. It is also possible to combine these if you want to reduce your input feature space to:
01100
There are other things to consider, as Scott has said in the comments:
How much data do you have? Maybe the feature space will be too big this way, I can't be sure. If you have enough data, then the model will intuitively learn the general distances (not actually, but intrinsically in the data) between datapoints.
If you have enough data, you might even be able to accurately predict between two points you don't have data for!
If it does come down to not having enough data, then finding representative features of the journey will come into use, ie. length of journey, shape of the journey, elevation travelled etc. Also a metric for distance travelled from the origin could be useful.
Best of luck!
I would be inclined to lean toward individual models. This is because, for a given position along a given route and a constant speed, the ETA is a deterministic function of time. If one moves monotonically closer to the target along the route, it is also a deterministic function of distance to target. Thus, there is no information to transfer from one route to the next, i.e. "lumping" their parameters offers no a priori benefit. This is assuming, of course, that you have several "trips" worth of data along each route (i.e. (distance, speed) collected once per minute, or some such). If you have only, say, one datum per route then lumping the parameters is a must. However, in such a low-data scenario, I believe that including a dummy variable for "which route" would ultimately be fruitless, since that would introduce a number of parameters that rivals the size of your dataset.
As a side note, NEITHER of the models you describe could handle new routes. I would be inclined to build an individual model per route, data quantity permitting, and a single model neglecting the route identity entirely just for handling new routes, until sufficient data is available to build a model for that route.
I have some data that tells me the amount of hours water is available for particular towns.
You can see it here
I want to use train a Multilayer Perceptron based on that data, to take a set of coordinates and indicate the approximate number of hours for which that coordinate will have water.
Does this make sense?
If so, am I correct in saying, there has to be two input layers? One for lat and one for long. And the output layer should be the number of hours.
Would love some guidance.
I would solve that differently:
Just create an ArrayList of WaterInfo:
WaterInfo contains lat,lon, waterHours.
Then for a given coordinate search the closest WaterInfo in the list.
Since you have not many elements, just do a brute force search, to find the closest.
You further can optimize, to find the three closest WaterInfo points, and calculate the weithted average of WaterHours. As weight you use the air distance from current position to Waterinfo position.
To answer your question:
"Does this makes sense"?
From the goal to get a working solution: NO!
Ask yourself, why do you want to use MLP for this task.
Further i doubt that using two layers for lat / long makes sense.
A coordinate (lat/lon) is one point on the world, so that should be one layer in the model. You can convert the lat/lon coord to a cell identifier: Span a grid over Brazil; with cell width 10 or 50km; now convert a lat/long coordinate to a cellId: Like E4 on a chess board, you will calculate one integer value representing the cell. (There are other solutions to get an unique number, too, choose one you like)
Now you have a modell geoCellID -> waterHours, which better represents the real world situation.
I need use the data of this site: http://www.navcen.uscg.gov/?Do=gpsArchives&path=2012
to develop a small software that plot a chart about satellate availability, something like this: http://i.stack.imgur.com/X0iGL.jpg
The user must set a day, a latitude/longitude position and a time zone, then my application must plot the satellate availability for 7 days (from user day) to choose the best day.
I'm not a GPS expert so I don't know which and how use the data from almanac to make the plot.
Any idea?
If you're good at Matlab you could try using this
This program calculate GPS visible satellites by exerting terrain for high accuracy prediction.
inputs:
Coordinate of Station on earth.
GPS Almanac file.
Terrain Data in "txt" format(just for DSM calculation).
Or you could go through the gpstk code and refer to how ComputeStationSatelliteVisibility l is implemented
The almanac contains the elipse parameters that describe the curve of one satellite around the earth. Using these params you could determine where the sats are positioned for a specific time and position.
Assume a visibility of 170° of the sky: 5° are hidden by houses or mountains at the horizon.
Refer to :
http://www.navcen.uscg.gov/?pageName=gpsAlmanacs
and
http://www.navcen.uscg.gov/pdf/gps/Programmatically%20Accessing.pdf
So I'm trying to solve a problem with Bayesian networking. I know the conditional probabilities of some event, say that it will rain. Suppose that I measure (boolean) values from each of four sensors (A1 - A4). I know the probability that of rain and I know the probability of rain given the measurements on each of the sensors.
Now I add in a new twist. A4 is no longer available, but B1 and B2 are (they are also boolean sensors). I know the conditional probabilities of both B1 and B2 given the measurement of A4. How do I incorporate those probabilities into my Bayesian network to replace the lost data from A4?
Your problem fits perfectly to Multi-Entity Bayesian Networks (MEBN). This is an extension to standard BN using First Order Logic (FOL). It basically allows nodes to be added and/or removed based on the specific situation at hand. You define a template for creating BN on the fly, based on the current knwoledge available.
There are several papers on it available on the Web. A classic reference to this work is "Multi-Entity Bayesian Networks Without Multi-Tears".
We have implemented MEBN inside UnBBayes. You can get a copy of it by following the instructions # http://sourceforge.net/p/unbbayes/discussion/156015/thread/cb2e0887/. An example can be seen in the paper "Probabilistic Ontology and Knowledge Fusion for Procurement Fraud Detection in Brazil" # http://link.springer.com/chapter/10.1007/978-3-642-35975-0_2.
If you are interested in it, I can give you more pointers later on.
Cheers,
Rommel
I want to detect the best rototraslation matrix between two set of points.
The second set of points is the same of the first, but rotated, traslated and affecteb by noise.
I tried to use least squared method by obviously the solution is usually similar to a rotation matrix, but with incompatible structure (for example, where i should get a value that represents the cosine of an angle i could get a value >1).
I've searched for the Constrained Least Squared method but it seems to me that the constrains of a rototraslation matrix cannot be expressed in this form.
In this PDF i've stated the problem more formally:
http://dl.dropbox.com/u/3185608/minquad_en.pdf
Thank you for the help.
The short answer: What you will need here is "Principal Component Analysis".
Apply this to both sets of points centered at their respective centers of mass. The PCA will effectively give you a rotation matrix for each aligned to the data set principal components. Multiplying the inverse matrix of the original set by the new rotation will give you a matrix that takes the old (centered) set to the new. Inverse translations and translations can similarly be applied to the rotation to create a homogeneous matrix that maps the one set to the other.
The book PRINCE, Simon JD. Computer vision: models, learning, and inference. Cambridge University Press, 2012.
gives, in Appendix "B.4 Reparameterization", some info about how to constrain a matrix to be a rotation matrix.
It seems to me that your problem has also a solution based on SVD: see the Kabsch algorithm also described by Olga Sorkine-Hornung and Michael Rabinovich in
Least-Squares Rigid Motion Using SVD and, more practically, by Nghia Kien Ho in FINDING OPTIMAL ROTATION AND TRANSLATION BETWEEN CORRESPONDING 3D POINTS.