Question

public final BufferedImage filter(BufferedImage src, BufferedImage dst)

Transforms the source BufferedImage and stores the results
in the destination BufferedImage.
If the color models for the two images do not match, a color
conversion into the destination color model is performed.
If the destination image is null, a BufferedImage is created with the source
ColorModel.
The coordinates of the rectangle returned by
getBounds2D(BufferedImage) are not necessarily the same as the coordinates of the
BufferedImage returned by this method. If the upper-left corner coordinates
of the rectangle are negative then this part of the rectangle is not drawn. If the
upper-left corner coordinates of the rectangle are positive then the filtered image is
drawn at that position in the destination BufferedImage.

I had the following code on Java 1.6:

//Make image always std_height tall
double scaleAmount = (double) std_height / (double) characterImage.getHeight();
AffineTransform tx = new AffineTransform();
tx.scale(scaleAmount, scaleAmount);
AffineTransformOp op = new AffineTransformOp(tx, AffineTransformOp.TYPE_BILINEAR);
characterImage = op.filter(characterImage, null);

In Android, i’m using Matrix instead of AffineTransform:

//Make image always std_width wide
float scaleAmount = (float) std_width / (float) characterImage.getWidth();
//AffineTransform tx = new AffineTransform();
//tx.scale(scaleAmount, scaleAmount);
Matrix mx = new Matrix();
mx.setScale(scaleAmount, scaleAmount);
//AffineTransformOp op = new AffineTransformOp(tx, AffineTransformOp.TYPE_BILINEAR); //Can't use this on Android
//characterImage = op.filter(characterImage, null); //Can't use this on Android

My problem is with the last two commented lines. Can i do something like it on Android? Thanks.

Answer 1

You can do something like

Matrix m = new Matrix();
matrix.postScale(scaleAmount, scaleAmount);
Bitmap b = Bitmap.createBitmap(original, 0, 0, original.getWidth(), original.getHeight(), m, true);

(Note: BTW, Oracle JavaSE AffineTransformOp#filter has a bug when the output BufferedImage is null, it creates a BufferedImage but with default colorspace, not the one of the input BufferedImage.)