As a computation unit for processing images, it could resize image to predefined size.
It will not stretch or compress the content of image. However, to fit the new size, it crops or pads pixels. When it crops image, it performs a center-crop; when it pads pixels, it performs a zero-padding.
ResizeWithCropOrPadOp(int targetHeight, int targetWidth)
Creates a ResizeWithCropOrPadOp which could crop/pad images to specified size.
|
TensorImage |
apply(TensorImage image)
Applies the defined resizing with cropping or/and padding on given image and returns the
result.
|
int |
getOutputImageHeight(int inputImageHeight, int inputImageWidth)
Computes the height of the expected output image when input image size is given.
|
int |
getOutputImageWidth(int inputImageHeight, int inputImageWidth)
Computes the width of the expected output image when input image size is given.
|
PointF |
inverseTransform(PointF point, int inputImageHeight, int inputImageWidth)
Transforms a point from coordinates system of the result image back to the one of the input
image.
|
Creates a ResizeWithCropOrPadOp which could crop/pad images to specified size. It adopts center-crop and zero-padding.
targetHeight | The expected height of cropped/padded image. |
---|---|
targetWidth | The expected width of cropped/padded image. |
Applies the defined resizing with cropping or/and padding on given image and returns the result.
Note: the content of input image
will change, and image
is the same instance
with the output.
image | input image. |
---|
Computes the height of the expected output image when input image size is given.
inputImageHeight | |
---|---|
inputImageWidth |
Computes the width of the expected output image when input image size is given.
inputImageHeight | |
---|---|
inputImageWidth |
Transforms a point from coordinates system of the result image back to the one of the input image.
point | the point from the result coordinates system. |
---|---|
inputImageHeight | the height of input image. |
inputImageWidth | the width of input image. |