Fingerprint Image Binarization

Question

I have an finger print scanner application which takes finger image data from device.

Now I am trying to binarize the image.

I am Using the Otsu's algorithm to binarize the image i.e. pixel's values either 0 or 255.

Threshold is calculated around 160 using same algorithm. Here is my code:

public static byte[][] binarizeImage(BufferedImage bfImage){
    final int THRESHOLD = 160;
    int height = bfImage.getHeight();
    int width = bfImage.getWidth();
    byte[][] image = new byte[width][height];

    for(int i=0; i<width; i++){
        for(int j=0; j<height; j++){
            Color c = new Color(bfImage.getRGB(i,j));
            int red = c.getRed();
            int green = c.getGreen();
            int blue = c.getBlue();
            if(red<THRESHOLD && green<THRESHOLD && blue<THRESHOLD){
                image[i][j] = 1;
            }else{
                image[i][j] = 0;
            }
        }
    }
    return image;
}

But the resulting image is not of the desired output.

Could anyone help me with this.

Answer 1

Otsu method isn't good for fingerprint images. Try to use this filters below:

• Bradley Local Threshold
• Bernsen Threshold.
• Maximum Entropy Threshold.

You will find here: http://code.google.com/p/catalano-framework/

Example:

FastBitmap fb = new FastBitmap(bufferedImage);
fb.toGrayscale();

BradleyLocalThreshold b = new BradleyLocalThreshold();
b.applyInPlace(fb);

bufferedImage = fb.toBufferedImage();