The basic idea is to apply a AlphaComposite
mask over the original image which has been filled with a LinearGradientPaint
So, we start by loading the original image...
BufferedImage original = ImageIO.read(new File("/an/image/somewhere"));
We then create a masking image of the same size...
BufferedImage alphaMask = new BufferedImage(original.getWidth(), original.getHeight(), BufferedImage.TYPE_INT_ARGB);
We then fill the masking image with a LinearGradientPaint
...
Graphics2D g2d = alphaMask.createGraphics();
LinearGradientPaint lgp = new LinearGradientPaint(
new Point(0, 0),
new Point(alphaMask.getWidth(), 0),
new float[]{0, 1},
new Color[]{new Color(0, 0, 0, 255), new Color(0, 0, 0 , 0)});
g2d.setPaint(lgp);
g2d.fillRect(0, 0, alphaMask.getWidth(), alphaMask.getHeight());
g2d.dispose();
What's important here is, we don't actually care about the physical color, only it's alpha property, as this will determine how the two images are masked together...
Then, we apply the mask...
BufferedImage faded = applyMask(original, alphaMask, AlphaComposite.DST_IN);
Which actually calls this utility method...
public static BufferedImage applyMask(BufferedImage sourceImage, BufferedImage maskImage, int method) {
BufferedImage maskedImage = null;
if (sourceImage != null) {
int width = maskImage.getWidth();
int height = maskImage.getHeight();
maskedImage = new BufferedImage(width, height, BufferedImage.TYPE_INT_ARGB);
Graphics2D mg = maskedImage.createGraphics();
int x = (width - sourceImage.getWidth()) / 2;
int y = (height - sourceImage.getHeight()) / 2;
mg.drawImage(sourceImage, x, y, null);
mg.setComposite(AlphaComposite.getInstance(method));
mg.drawImage(maskImage, 0, 0, null);
mg.dispose();
}
return maskedImage;
}
This basically uses a "destination in" AlphaComposite
to apply the mask onto the original image, which results in...
(original on the left, alpha on the right)
And just to proof the point, I changed the background color of the frame's content pane to RED