Hi,

I am newbie to image processing and i am using this framework for first time.

I am getting null pointer exception when your code try to access gray of the image
Exception in thread "main" java.lang.NullPointerException at Catalano.Imaging.FastBitmap.getGray(FastBitmap.java:942) at Catalano.Imaging.Tools.ObjectiveFidelity.getMSE(ObjectiveFidelity.java:122) at qualityanalysis.QualityAnalysis.calculate(QualityAnalysis.java:44) at qualityanalysis.QualityAnalysis.main(QualityAnalysis.java:105)

I am using simple code mentioned here in the answer.

hello where is the main of gwo file?

Hi,@DiegoCatalano
thanks a lot and appreciate you for the useful library but have you ever tested SingleScaleRetinex and other samples like MultiScaleRetinex ???

• Catalano Framework: 1.4
• Android 4.4.4

Hello Diego,
if i run this sample code:

FastGraphics graph = new FastGraphics(image);
graph.setColor(180, 255, 50);
graph.DrawLine(new IntPoint(10, 10), new IntPoint(10, 110));
_imgView.setImageBitmap(image.toBitmap());

I expect a vertical line to be drawn but an horizontal one is drawn instead (is IntPoint(x, y) right?)

I have another issue, maybe related to the above mentioned (major issues in the pixel coordinate management?), if I call the DrawCircle this way:

_graph.DrawCircle(new IntPoint(269, 15), 5)

I got this exception:
java.lang.ArrayIndexOutOfBoundsException: length=76800; index=87695
at Catalano.Imaging.FastBitmap.setRGB(FastBitmap.java:438)
at Catalano.Imaging.FastGraphics.DrawCircle(FastGraphics.java:97)
at Catalano.Imaging.FastGraphics.DrawCircle(FastGraphics.java:81)

The frame size is 320 x 240 (width x height) so, theoretically I should be able to draw a 5 pixels radius circle at 269,15 (x,y).

Anyway, thank you for this great framework.
Claudio

Hi @DiegoCatalano,

       While i captured image using android Indent and after converting to mutable bitmap image with **_Bitmap.Config.ARGB_8888_** and passing to Fastbitmap  the following error arises,

E/AndroidRuntime: FATAL EXCEPTION: main
Process: catalano.imaging, PID: 3401
java.lang.OutOfMemoryError: Failed to allocate a 31961100 byte allocation with 12540956 free bytes and 11MB until OOM
at catalano.imaging.FastBitmap.refresh(FastBitmap.java:179)
at catalano.imaging.FastBitmap.(FastBitmap.java:82)

Please any one give solution to solve this,

Thanking You !!

I was reading the filters and I found something odd. Every part of the Blend filter fastBitmap is updated but in line 357 overlay is updated, is it a typo?

Blend.java#L357

Dear Catalano,

It seems this is the only way to reach you. I found HMM models on Accord net framework very useful. Do you have any plans to port HMM classifier to Java and Android? If you, when do you expect to finish?

Hey,

i tried using the concurrent HSLFilter in an Android application. It didn't work at all, so i tried debugging the app:

Previously the Share-Object is created like this here:
t[i] = new Thread(new Run(new Share(fastBitmap, startX, startX + part)));

Since the Share object is initialized without the endWidth parameter, it is 0, which will cause the algorithm to never execute the inner for-loop.

Hope this can be fixed in the next release.

Anyone can give me some sample programs or codes that can get the hu moment of images using this library .I really need your help. I will appreciate your help. Thanks in advance! this is sample images that used in my project

HI Diego,

I use your nice Framework, but now as I switch to android I get a exception during fastBitmap.toBitmap().
Exception is:
java.lang.IllegalStateException: Can't call setPixels() on a recycled bitmap
07-01 09:42:18.942 12791 13040 E AndroidRuntime: at android.graphics.Bitmap.checkRecycled(Bitmap.java:243)
07-01 09:42:18.942 12791 13040 E AndroidRuntime: at android.graphics.Bitmap.setPixels(Bitmap.java:1250)
07-01 09:42:18.942 12791 13040 E AndroidRuntime: at Catalano.Imaging.FastBitmap.toBitmap(FastBitmap.java:854)

Can you help me? Maybe it´s a problem because i recycle the Bitmap to fast?

I am looking for a active contour algorithm to find object contours in images.
Can you tell me how to use you snake implementation?
Do you have an snipet as an example?

Thanks

hi, Catalano, fft process returns blank (full wihite) image. where i m doing wrong? i attached code and input, output images.

FastBitmap fb = new FastBitmap(imagePath);

Grayscale grayscale = new Grayscale();
grayscale.applyInPlace(fb);
fb.saveAsJPG(path + "grayscale_" + FilenameUtils.getBaseName(imagePath) + ".jpg");

FourierTransform ft = new FourierTransform(fb);
ft.Forward();
fb = ft.toFastBitmap();
fb.saveAsJPG(path + "fft_" + FilenameUtils.getBaseName(imagePath) + ".jpg");

FrequencyFilter ff = new FrequencyFilter(0, 60);
ff.ApplyInPlace(ft);
fb = ft.toFastBitmap();
fb.saveAsJPG(path + "freq_" + FilenameUtils.getBaseName(imagePath) + ".jpg");

Hello,

Does the BlobDetection create blobs having their x and y coordinates rotated? Is this a feature or a bug?

From the BlobDetection codes, it seems, strangely enough, that x is used iterate height and y is used to iterate width when the blobs are created. What is the reason for this?

Good work on the framework :)

Is there a way to avoid OutOfMemory exception when editing images with high resolution?

I have created a package and placed the files in it, Android Studio complains that it cannot resolve symbol Catalano, even though autocomplete suggest it just fine. A maven or jar to import would be much appreciated to avoid this hassle with AS importing.

Are there any plans for an red eye removal algorithm?
I cannot find any good solution for this in java.

Typically, a feature descriptor (HOG descriptor) converts an image of size width x height x 3 (channels ) to a feature vector / array of length n.
The ProcessImage method in the file HistogramOfOrientedGradients.java has desription "Process image looking for corners" and returns "list of found corners (X-Y coordinates)".
Can You explain what is doing this class and how to interpret the return values?
Thank You.

Hello,

The ImageStatistics class seems to throw&print exception stack traces unnecessarily (in several locations). They are all printed on System.out.
Example:
try {
throw new Exception("Histogram red null");
} catch (Exception e) {
e.printStackTrace();
}

You could try adding logging (JDK/log4j/...) instead?

Br,
Saku

Hi,

I just found this pixelBrightnessDiff param is unused here

Maybe I can help create PR for this one

Hi Diego,

I am trying to use the FAKE HDR in Android but I get this error on using it. Can you help in providing pointers as to where to look.

Thanks!

Hi DiegoCatalano,

                 I am now trying to use your updated Catalano-Framework for android image processing, I am getting issues of some missing sub files. Can you provide Jar Files or Gradle for Core,Fuzzy,Math and other libraries. 

Hi Diego,

want to report I got this crash

once I want implement the wavelet transformation

what's wrong with my code? the height and width looks fine?

Hi Catalano,you have a lot of filters available in your library that i am really confused which one to use,can you put up images of every filter taking a single photo and applying every one on that so that everyone can easily get to know what each filter does to my photo.It would be a great help.
And please add the vignette effect if possible.
Thanks

I am processing an Image to make it Tesseract-OCR friendly using Catalano framework. Thresholding is working fine but when I try to UnsharpMasking or resizing I get an error Couldn't correct orientation: java.lang.IllegalStateException: Can't call setPixels() on a recycled bitmap even though I have set BitmapFactory options inmutable to true & minsdk=11.

Code snippet:
FastBitmap fb = new FastBitmap(bitmap);
Grayscale grayScale = new Grayscale();
grayScale.applyInPlace(fb);

        switch (thresholdFilterName) {
        case "Threshold":  Threshold t = new Threshold(100);
        t.applyInPlace(fb);
        break;

        case "BradleyLocalThreshold":  BradleyLocalThreshold blt = new BradleyLocalThreshold();
        blt.applyInPlace(fb);
        break;

        case "SauvolaThreshold":  SauvolaThreshold st =new SauvolaThreshold();
        st.applyInPlace(fb);
        break;

        case "NiblackThreshold":  NiblackThreshold nt =new NiblackThreshold();
        nt.applyInPlace(fb);
        break;

        case "WolfJoulionThreshold": WolfJoulionThreshold wjt =new WolfJoulionThreshold();
        wjt.applyInPlace(fb);
        break;

        default: WolfJoulionThreshold wjtD =new WolfJoulionThreshold();
        wjtD.applyInPlace(fb);
        break;

        }


        UnsharpMasking um= new UnsharpMasking();
        um.applyInPlace(fb);
    // exception is thrown here.

Another thing is that BradleyLocalThreshold gives best result in lighting conditions where there is shadow on image but the accuracy of OCR engine is poor with it, where as accuracy with HysteresisThreshold is quiet high but it doesn't work well in lighting conditions. Is there any way we have best of both algorithms.

Hi, I really like your work on image processing.
It would be great if you can add the filters same to those in Instagram (Amaro,Lomo-fi,etc)
It would be a great help.
thanks

Hi,

Can someone tell me how to perform perspective transformation using this framework?

Please refer the images in this thread regarding perspective transformation
http://answers.opencv.org/question/102464/android-opencv-perspective-transformation-not-working-as-expected/

Anyone can give me some sample programs or codes that can get the pixels of an image. Total pixels, pixels per fold, how to fold an image, cut the image excess space of the image. I really need your help. I will appreciate your help. Thanks in advance!

Hi,

do you plan to integrate a green screen (color key) masking algorithm?
Or do you have any advice for me ho i can achieve something like this?

Thanks
Sebastian

What if I create fork and add ability to build tihs framework with maven?
It is very useful and why not to publish it in maven central?

Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 44000
 at Catalano.Imaging.FastBitmap.setGray(FastBitmap.java:813)
 at Catalano.Imaging.Filters.Erosion.applyInPlace(Erosion.java:94)
 at Catalano.Imaging.Filters.Opening.ApplyInPlace(Opening.java:80)
 at Catalano.Imaging.Filters.Opening.applyInPlace(Opening.java:63)
...

Filters applied before:

• Grayscale
  • Grayscale.Algorithm.Lightness
• Crop
  • x = 50
  • y = 0
  • width = 440
  • height = 100
• Threshold
  • value = 30
  • invert = true

We wanted to apply an Opening filter with radius = 5.

It works fine with the jars from codeproject but neither with the source from their nor with the source from github.

Hi Folks,
This is one of the most useful image processing libraries for Android. And what for bytes (lean and mean). Presently I need to increase the dpi(resolution) of an image. Should I use ResizeBicubic?

I found a bug in the threshold class, when it is not inverted a value greather or equals than the specified is white, but when it is inverted a value less than the specified is black, painting in both cases the same images. Just need to change the condicional from less than to greather or equals to.

Greetings

Hello, is it possible to display mjpeg stream using this framework?

Hello,

I've been running few algrothms against my material (>50MB images). Do you have in your roadmap a slot to optimize the Convolution process to use direct pixel byte/int arrays (produced by FastBitMap.getGrayData() and FastBitMap.getRGBData()) instead of using copy.getGray/Red/Green/Blue methods?

This might speed up things (blurs, edgedetectors) considerably.

BR,
Saku

Hi there,
I want to compute Zernike Moment. Have referred to your code for computing Zernike moments, however what input shall be given to the ZernikeMoments() function is unclear to me.

viz.
/**
* Compute the set of Zernike's moments up to the specified order.
* Also compute the width, height, and centroid of the shape.
* @param order Order.
* @param x X axis coordinates.
* @param y Y axis coordinates.
* @param npoints Indicates the total number of points in the digitized shape.
* @return Zernike`s moments.
*/
public static ComplexNumber[] ZernikeMoments(int order, double[] x, double[] y, int npoints){
}

What are the X axis coordinates and Y axis coordinates?
From what I understand, if it's a coordinate, it should be simple value double x, double y, and not any array.
Also for computing moment of image, function should take input as a image as maybe BufferedImage or 2D image array[][].

Sorry for being so naive. I am image processing newbie, and weak in math.

Thank you.

Hi, I am trying to implement the Homomorphic filter to some images, but it doesn't seem to work correctly for some reason. Is there any sample code for this?

Also I am not aware about the min,max range of frequencies that are used. What do they represent?

I have tried both RGB and Grayscale images. Is there some theory I am missing and can't get that to work?

Hi,
A question about gaussian.java
The equation of Function2D is written :
Math.exp(-(x * x + y * y) / (2 * sqrSigma))
But according to wikipedia
Shouldn't it be:
Math.exp(-(x * x + y * y) / (2 * sqrSigma))/(2*Math.PI*sqrSigma)

Thank you for the framework, Vincent

Do you have face recognition solution?
I cannot find any good solution for this in java.

Hello,
I am trying to find a way to remove horizontala and vertical lines from scanned Documents, to have better results with Tessercat. I tried with Hough, but I am getting lost...
It would be great to have some hints!

Thnaks in avance

Hi all,
I'm playing with this framework with good results. In particular I'm using the clahe filter to enhance my images. Unfortunately my images are satellite data that comes as 16bits unsigned short (from 0 to 65535 per band) or 32bit float (from 0 to 1).

So I had to manually convert the image to 8 bit before applying the filter. Do you have any plan in the future to handle directly those type of images?
I've tried looking into the code myself but it appears not to be an easy task.

Thank you for all your work!

Hello Diego,
using Sauvola Threshold filter i found that process gets stuck when the SeparableConvolution is executed. I'm applying it on an image with a resolution of about 2500x1500 and the process never ends because all the iterations SeparableConvolution performs.

Was this filter designed to handle only small resolution images? Or can this be improved for bigger resolutions?

Thanks

I did try to look for correlation function in the Catalano Framework, but I've not found anything that would return to me an array of matching point pairs.

Currently I am trying to use the code from http://www.codeproject.com/Articles/95453/Automatic-Image-Stitching-with-Accord-NET in my Java desktop app.

PS: haven't found RANSAC either :(

Hi,

I would like to know the functions to create 8-bit grayscale image from android bitmap and save as jpeg/jpg.

If I use android method to create image from bitmap, it is not saving images in 8-bit depth, since ALPHA_8 doesn't work in android

Thanks.

Sir, can you help me use your catalano`s framework? I am trying to get the total pixels of an image and cut all the excess space for me to get the total width and height of the used space of the image in an android studio project. Can you please give some sample codes for this? I would really appreciate your help.

Hello Sir,
In Catalano.Imaging.Filters, most methods alter
pixelsGray[xstrideX+ystrideY] = (byte) value using FastBitmap.setGray method.
How does pixelsGRAY affects this.bufferedImage in FastBitmap.toIcon/FastBitmap.saveAsJPG?

Also, I am having difficulties displaying images with DataBufferUShort after altering their FBM.pixelsGray[], may I ask if you have any advice to display/save the altered images?

Currently, Ive tried altering FastBitmap.java as follow
private void refresh() {
this.raster = getRaster();

	if (isGrayscale()){
		if (raster.getDataBuffer() instanceof DataBufferByte){
		    pixelsGRAY = ((DataBufferByte) raster.getDataBuffer().getDat
		    a();
		    }
	 else if (raster.getDataBuffer() instanceof DataBufferUShort){
		    pixelData = ((DataBufferUShort) raster.getDataBuffer().getDat
		    a();
		    ByteBuffer byteBuffer = ByteBuffer.allocate(pixelData.length*2);
		    byteBuffer.asShortBuffer().put(ShortBuffer.wrap(pixelData));
		    pixelsGray = byteBuffer.array();
		    
		    }
		}

}

The values in pixelsGray became negative when using invert(), however the this.bufferedImage remains the same.

Can you please give me an example how to calculate Zernike moments. I did not find any example in the sample folder.
I'm really need it.

I'm trying to use the framework to do some grid detection. So far, I found the BradleyLocalThreshold class, but I'm not sure how I should use it exactly. Also, there are some other things that I want to do like contour detection and Gaussian blurring, is there any general documentation or sample code on image processing that I can look at?

From my limited research, it looks like there used to be samples and documentation in the code.google.com project, but nothing is there anymore and I can't find anything in this repo.

Hi, thanks for porting Aforge library, btw I am not seeing document skew checking in the current version. can you add it?

http://www.aforgenet.com/framework/docs/html/18d8bd4d-b15e-8507-a480-b2fe6eaeab44.htm

Thanks

Hi Diego,

Just found out this class on AForge but I didn't found out that class on 1.6.1.1 version, is that same class as HistogramStretch?