Develop Reference

Detect eye ball

I am developing an app which will change the color of your eye. I need some help about detecting eye ball. Currently i have a selector that will be used to reduce ROI. It look like
this
So who we can detect eye ball from that selected region. I was thinking about changing the image to grayscale and then detect big black spot from that and then change color of it which will be the next step. i'll really appreciate any help.

Your way of thinking about a returning pattern is a good start. I am doing some work on a pattern recognition chair as well, so here is some help for your task:
using a grayscale is a good start btw ;)
There are some "facts" that are always applicable to a non pathologic eye:
the center is dark
left and right side surrounding dark ball are almost white (depends on how open the eye is)
do not forget: you have 2 eyes. link them together in some way (usually they are on an approximately horizontal line)
there is usually motion in the eyes while the other regions of the picture are relatively calm
Of course I cannot provide any code here, this would blast this whole post, but I hope I could help you in some way.

I found some link these guys detecting pupil of the eye. May b this will help you . See here and here

you could use the template matching method from open cv. template matching
this will help you find the eye in most of the cases.
Another solution would be to convert your image into an edge image with e.g. canny edge detector from opencv. and then search for this pattern with the template matcher. Using the edges makes you independent of the color. Using grayscale images will also facilitate the procedure.

Android: How to detect these objects in images? (Image included). Tried OpenCV and metaioSDK, but both are not working good enough

i have been working with object detection / recognition in images captured from an android device camera recently.
the object i am trying to detect are all kinds of buttons that look like this:
Picture of buttons
so far i have been trying with OpenCV and also with the metaio SDK. results:
OpenCV was always detecting something, but gave lots of false hits. also it is too much work to collect all the pictures for what i have in mind. i have tried three ways with OpenCV:
FeatureDetection (SURF, ORB and so on) -> was way too slow and not enough features on my objects.
Template Matching -> seems to only work when the template is exactly a part out of the scene image
Training classifiers -> this worked the best so far, but is too much work for my goal, and still gives too many false detections.
metaioSDK was working ok when i took my reference images (the icon part of each button) out of a picture like shown above, then printed the full image and pointed my android device camera at the printed picture. but when i tried with the real buttons (not a picture of them) then almost nothing got detected anymore. in the metaio documentation it is said that the reference images need to have lots of features and color differences and also should not only consist of white text. well, as you see my reference images are exactly the opposite from what they should be. but thats just how the buttons look ;)
so, my question would be: does any of you have a suggestion about what else i could try to detect and recognize each of those buttons when i point my android camera at them?

As a suggestion can you try the following approach:
Class-Specific Hough Forest for Object Detection
they provide a C code implementation. Compile and run it and see the results, then replace positive and negative training images with the ones you have according the following rules:
In a car you will need to define the following 3 areas:
target region (the image you provided is a good representation of a target region)
nearby working area (this area have information regarding you target relative location) I would recommend: area 3-5 times the target regions, around the target, can be a good working area
everything outside the above can be used as negative images
then,
Use "many" positive images (100-1000) at different viewing angles (-30 - +30 degrees) and various distances.
You will have to make assumptions at which viewing angles and distances your users will use the application. The more strict they are the better performance you will get. A simple "hint" camera overlay can give a good idea to people what you expect the working area to be.
Use few times (3-5) more different negative image set which includes pictures of things that might be in the camera but should not contribute any target position information.
Do not use big images, somewhere around 100-300px in width should be enough
Assemble the database, and modify the configuration file that the code comes with. Run the program, see if performance is OK for your needs.
The program will return a voting map cloud of the object you are looking fore. Add gaussian blur to it, and apply some threshold to it (you will have to make another assumption for this threshold value).
Extracted mask will define the area you are looking for. The size of the masked region can give you good estimate of the object scale. Given this information it will be much easier to select proper template and perform template matching.
(Also some thoughts) You can also try to do a small trick by using goodFeaturesToTrack function with the mask you got, to get a set of locations and compare them with the corresponding locations on a template. Constuct an SSD and solve it for rotation, scale and transition parameters, by mimizing alignment error (but not sure if this approach will work)

How can we draw any graphics with just only canvas?

There are many curious things in android.
The thing that I want to ask is about Drawing with canvas.
when I override ondraw function ,
and the android gives me canvas to draw something in device.
I would like to get answer how canvas draw image or some figures internally.
If they need to draw something, use surface flinger or openglES or core graphics or any
thing to swap image from back buffer to front buffer in device display.
I got some similar answers in this link Android GUI architecture - relation between Surface/view/window/canvas
But it's not really enough to understand..
Because I can't understand how they can initialize the device display and draw canvas in device display
I understand what relationship is in ViewRoot and View also surface.
Please let me know keys to draw something deeply about native layer.

I strongly suggest you load the API Demos app into Eclipse and look through the Graphics demos. There are a large number of them and the code should give you insights into solutions to your problem.
To load the API Demos app, in Eclipse click File > New > Other > Android Sample Project.
Then select your build target such as Android 4.2.
Then from the list of sample apps select the API Demos app.
And finally click Finish.
There are over 300 demos in the app, so it can be difficult to navigate. You want to look in the Graphics section.
It's a little unclear from your description exactly what your issue is, but the API Demos cover a lot of territory and you should be able to find some code to help you out.

Here is a list of some of the Graphics API Demo Java source files:
AlphaBitmap
AnimateDrawables
Arcs
BitmapDecode
BitmapMesh
BitmapPixels
CameraPreview
Clipping
ColorFilters
ColorMatrixSample
Compass
CreateBitmap
DensityActivity
FingerPaint
Layers
MeasureText
PathEffects
PathFillTypes
Patterns
Pictures
DrawPoints
PolyToPoly
Regions
RoundRects
ScaleToFit
SensorTest
SurfaceViewOverlay
WindowSurface
Sweep
TextAlign
TouchPaint
Typefaces
UnicodeChart
Vertices
Xfermodes

Drawing a Circular Timer (AndEngine)

I'm using AndEngine, and within that framework, I'd like to make a circular timer graphic. Specifically, I'd like to display the wait period for reuse of an ability. The idea is to dynamically fill the arc as the timer progresses. Something like this:
The java.awt.Graphics object has a fillArc() method that seems perfect for me. In fact, the above graphic was drawn using fillArc(50,5,100,100,75,-40). Great! Now here's the problem:
AndEngine doesn't use Graphics() objects, it uses its own Shape implementation (for OpenGL) and there's no defined "Circle" shape, much less a circle shape with a fillArc() method.
Possible Solutions and Their Respective Problems
Looking around for a solution I ran into "Drawing a circle using Andengine". That Q&A is not of much use to me as the only answer "Indeed, you can't directly draw a circle" offers two alternatives: (1) "Rendering a quad with a circle texture" - this won't work for me as I need to dynamically modify the circle to produce the arcfill; and (2) "Rendering a circle that's actually a circle of connected triangles." Maybe option two would work, but there's no guidance there as to how to do that.
I also ran into "Creating circle in android andengine by box2d?". I suspect someone may be tempted to say, you can simply create a circle like this:
Body circleBody = PhysicsFactory.createCircleBody(pWorld, pSprite,
BodyType.StaticBody, FixtureDef);
That really doesn't help me. I'm not looking to create a 2D physics body of a circle. I'm looking to display one.
Finally, I found this discussion, which is promising. In particular, there's a suggestion:
Use Canvas to draw [it] into a Bitmap, and load that Bitmap as a TextureSource.
Sounds reasonable, although I'm still unclear how to do that.
Update: My Cheating "Solution"
Rather than dwell on this, I decided to cheat (for the moment at least). I made a spritesheet that looks like this:
Rather than actually have the timer display the perfect fillArc(), I just pull the appropriate index of the sprite from the spritesheet based on rounding the proportion done (from 0 to 1) to the appropriate index on the spritesheet. Like this:
public void setTimer(float amount) {
this.setCurrentTileIndex(Math.round(amount * 20));
}
For my purposes, this actually works just fine--I'm using the timers over about 2 seconds, so you really don't see the lack of detail. But maybe I'll get around to replacing this with the "proper" solution if someone posts it. Also, maybe this spritesheet will be useful for someone doing the same thing. Here's the version using transparency instead of a green background. (So it's white on the white background of stackoverflow, but it's there):

There is a third solution that requires a single texture and a custom object. So it's a trade off between your solutions, where one requires a lot of triangles and the other one a texture memory.
You need only one image, i.e. the full circle in your "cheat sequence" above.
Create a custom object consisting of 8 triangles (one 'fully drawn' triangle will represent 45° each).
The progress determines:
How many of the triangles to draw. I.e.:
100% ==> 360° ==> 8 full triangles
50% ==> 180° ==> 4 full triangles
37.5% ==> 135° ==> 3 full triangles
25% ==> 90° ==> 2 full triangles
20% ==> 72° ==> 1 full triangle and one triangle with one vertex moved so that it represents the remaining 27° (== 72° - 45°).
If you ask me this is the coolest solution, since it can be applied to any texture. =)

How to project a picture onto a bubble surface programmatically

I want to implement an effect that, showing a picture on a bubble surface with the picture looks like surrounding the bubble surface. But I don't know how to do this kind of things...
I am doing it in Android platform, should I use OpenGL ES for this ? Or just some 2D transformation can achieve this effect ?
One more question is, I want to create many interesting graphics effects like the PhotoShop's various filter, is there any books/articles I can refer for this kind of things ? Is this kind of work belongs to the "Digital Image Processing" field or some other computer graphics related fields ?

Or just some 2D transformation
This effect is a nonlinear transformation, so doing through the 2D (linear) transformations being available will not work. You can do it using OpenGL by numerous ways. I'm currently thinking about some easy to understand way to convey, what you need to do. Basically you need to implement some kind of refraction or nonlinear radial warp.
Say p is the center of your bubble (in 2D), and r the position relative to p, then the undistorted picture is given by r+p. Now you want to distort it toward the edges. A parabolic distortion comes to mid, i.e. instead of r+p you'd show the pixel r + (|p|^2)*p/|p|