Using a Kinect sensor, I am attempting to write an algorithm to detect a clenched fist. I am trying to achieve this by calculating the area occupied by the hand (since clenched fist area < non-clenched fist area).
Here's what I have so far:
Depth information per pixel (from the DepthStream)
Location of the Hand (from the SkeletonStream)
I am having trouble figuring out how to get the depth data that corresponds to the hand. It's easy to get the depth data at the exact location that the Kinect gives for the hand, but I don't know how to get all the depth data for the hand. Any suggestions, pseudocode, and/or link to tutorials would help.
There are events from KinectInteractions that detect whenever the fist is in gripped mode or released mode as it's used for the KinectScrollViewer in the KinectRegion:
The HandPointGrip event
The HandPointGripReleased event
Also this might be a duplicate of this post .
I am not experienced in this field but found this in my search. See if it helps you: https://groups.google.com/forum/#!topic/openni-dev/Kj2JL6K0PBw
Related
I use a Kinect SDK V2 sensor to capture point cloud.The output of my program is a double array that contains X-Y-Z position points for every frame(0,0,0 is the center of the sensor). Now i would like to visualize the data for example for one frame. Is any fast/easy way just to visualize the data?Any .dll?
P.S:I tried with Unity but it was to difficult and time consuming for me.I thought that code does not need it because question it too general(if code would be helpful i could upload parts).
Thank you
I'm using Oxyplot HeatMapSeries for representing some graphical data.
For a new application I need to represent the data with isosurfaces, something looking like this:
Some ideas around this:
I know the ContourSeries can do the isolines, but I can't find any option that allows me to fill the gaps between lines. Does this option exists?
I know the HeatMapSeries can be shown under the contourSeries so I can get a similar result but it does not fit our needs. .
Another option wolud be limiting the HeatMapSeries colours and eliminate the interpolation. Is this possible?
If anyone has another approach to the solution I will hear it!
Thanks in advance!
I'm evaluating whether Oxyplot will meet my needs and this question interests me... from looking at the ContourSeries source code, it appears to be only for finding and rendering the contour lines, but not filling the area between the lines. Looking at AreaSeries, I don't think you could just feed it contours because it is expecting two sets of points which when the ends are connected create a simple closed polygon. The best guess I have is "rasterizing" your data so that you round each data point to the nearest contour level, then plot the heatmap of that rasterized data under the contour. The ContourSeries appears to calculate a level step that does 20 levels across the data by default.
My shortcut for doing the rasterizing based on a step value is to divide the data by the level step you want, then truncate the number with Math.Floor.
Looking at HeatMapSeries, it looks like you can possibly try to turn interpolation off, use a HeatMapRenderMethod.Rectangles render method, or supply a LinearColorAxis with fewer steps and let the rendering do the rasterization perhaps? The Palettes available for a LinearColorAxis can be seen in the OxyPalettes source: BlueWhiteRed31, Hot64, Hue64, BlackWhiteRed, BlueWhiteRed, Cool, Gray, Hot, Hue, HueDistinct, Jet, and Rainbow.
I'm not currently in a position to run OxyPlot to test things, but I figured I would share what I could glean from the source code and limited documentation.
Is it possible to Extract any Shape that's in front of an image?
let's say we have an image of two objects 1 in front, the other is behind and a blank or transparent background.
can we extract the one in front and place it in a new image?
can this be done by detecting edge of frontal shape and then crop it?
This article is doing something near to my question :
Cropping Particular Region In Image Using C#
but i want to do it fully automated.
any help would be highly appreciated.
Thanks in advance.
I think you cannot do this fully automated; however, there are maybe some semi-automated ways, at least, you need some prior information such as how far your object can be placed. Here are some of my suggestions.
First way(you have experience in implementing academic papers, you have some prior information about depth of object's place),
Download a "scene - depth images database" from internet
Get the average value of database
Query K-Nearest-Neighbors of input image according to GISt of scene[1]
Apply SIFT flow to align database scenes according to input scene
Infer the depth
Remove a certain range from image.
It's possible to infer rough depth map of an input image. By using this, you'll try to infer depth map of input image then remove the certain depth range which includes your object. You can check the paper[2] for more detailed explanation.
Example Depth Map from http://www.the.me/wp-content/uploads/2013/09/z-depth_map_expanding_exif_more_powerful_post-processing_2n.jpg
Second way(assumption: human input is allowed at the end of algorithm.),
- Segment the image(you can find state of the art algorithm with a little search)
- Select the contour that you want to remove.
Example Segmented Image from http://vision.ece.ucsb.edu/segmentation/edgeflow/images/garden_edge.gif
References:
[1]Aude Oliva
Gist of the Scene
[2]Karsch, K.; Liu, C.; Kang, S.B.
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2014.
OpenCV gives option for to extract contours of the objects. So convert your image to gray scale and given to the open CV , detect all the contours in your images. and from that select contours which will be required for your requirement.
Since your project is on C# you can take a look of Emgu CV which is a cross platform .Net wrapper for OpenCV. Pl. refer to the below url where you can download the examples for Emgu CV.
http://sourceforge.net/projects/emguexample/?source=recommended
I am trying to figure out a way of getting Sikuli's image recognition to use within C#. I don't want to use Sikuli itself because its scripting language is a little slow, and because I really don't want to introduce a java bridge in the middle of my .NET C# app.
So, I have a bitmap which represents an area of my screen (I will call this region BUTTON1). The screen layout may have changed slightly, or the screen may have been moved on the desktop -- so I can't use a direct position. I have to first find where the current position of BUTTON1 is within the live screen. (I tried to post pictures of this, but I guess I can't because I am a new user... I hope the description makes it clear...)
I think that Sikuli is using OpenCV under the covers. Since it is open source, I guess I could reverse engineer it, and figure out how to do what they are doing in OpenCV, implementing it in Emgu.CV instead -- but my Java isn't very strong.
I looked for examples showing this, but all of the examples are either extremely simple (ie, how to recognize a stop sign) or very complex (ie how to do facial recognition)... and maybe I am just dense, but I can't seem to make the jump in logic of how to do this.
Also I worry that all of the various image manipulation routines are actually processor intensive, and I really want this as lightweight as possible (in reality I might have lots of buttons and fields I am trying to find on a screen...)
So, the way I am thinking about doing this instead is:
A) Convert the bitmaps to byte arrays and do brute force search. (I know how to do that part). And then
B) Use the byte array position that I found to calculate its screen position (I'm really not completely sure how I do this) instead of using the image processing stuff.
Is that completely crazy? Does anyone have a simple example of how one could use Aforge.Net or Emgu.CV to do this? (Or how to flesh out step B above...?)
Thanks!
Generally speaking, it sounds like you want basic object recognition. I don't have any experience with SIKULI, but there are a number of ways to do object recognition (Edge based template matching, etc.). That being said you might be able to go with just straight histogram matching.
http://www.codeproject.com/KB/GDI-plus/Image_Processing_Lab.aspx
That page should show you how to use AForge.net to get the histogram of an image. You would just do a brute force search using something like this:
Bitmap ImageSearchingWithin=new Bitmap("Location of image"); //or just load from a screenshot or whatever
for (int x = 0; x < ImageSearchingWithin.Width - WidthOfImageSearchingFor; ++x)
{
for (int y = 0; y < ImageSearchingWithin.Height - HeightOfImageSearchingFor; ++y)
{
Bitmap MySmallViewOfImage = ImageSearchingWithin.Clone(new Rectangle(x, y, WidthOfImageSearchingFor, HeightOfImageSearchingFor), System.Drawing.Imaging.PixelFormat.Format24bppRgb);
}
}
And then compare the newly created bitmap's histogram to the one that you calculated of the original image (whatever area is the closest in terms of matching is what you would select as being the region of BUTTON1). It's not the most elegant solution but it might work for your needs. Otherwise you get into more difficult techniques (of course I could be forgetting something at the moment that might be simpler).
How can you filter pornography in a video/picture?
Is there a library I can use? Commercial or open source is OK.
For Video:
You would analyze the sound waves from the video. If there's lot of grunting and moans, find a way to recognize those patterns and label it as porn. However, you might get false tags for women's tennis match.
For Image:
Build a simulation model of a human brain which signals arousal from viewing erotic pictures. Some paramters would be excessive curves, and image recognition of the human anatomy. This is far more difficult.
If it is not a real-time, you could use Amazon's Mechanical Turk to do the work. You'll have to pay the workforce, but it's a very cheap way to get menial jobs like this done.
See, Porn Detection is not a very easy.
But there is surely a range of color that identifies the human skin. And you need to make a threshold of the color to identify the picture as porn. Say you take it as 70 %, so you go through all the pixel on the current screen and determine whether 70% of the pixels comes within the range of human skin color.
There is already few algorithm prepared for Face detection like :
http://www.codeproject.com/KB/cs/Face_Detection_processing.aspx
I hope this would help you to find a way on your point.