Overview

Gaussian pyramid generator takes one input image and fills the output image pyramid with downscaled versions of the input.

Input	Output

Implementation

The function is implemented by generating the Gaussian pyramid from the base (level 0) to coarser levels.

If the input image actually wraps the first level of the image pyramid, nothing is done for this level. If not, the input image contents will be copied to the first image pyramid level.

The coarser levels are generated by taking the previous level, convolving it using a clamp boundary condition with the following kernel:

\[ k = \begin{bmatrix} 1 \\ 4 \\ 6 \\ 4 \\ 1 \end{bmatrix} \cdot \begin{bmatrix} 1 & 4 & 6 & 4 & 1 \end{bmatrix} \]

Because only 2x downscaling is supported, the result is then downsampled by keeping all pixels with even coordinates.

The algorithm repeats until all levels are generated.

Usage

Initialization phase
1. Include the header that defines the Gaussian pyramid generator function.
  #include <vpi/algo/GaussianPyramidGenerator.h>
2. Define the stream on which the algorithm will be executed and the input image.
  VPIStream stream = /*...*/;
  
  VPIImage input = /*...*/;
3. Create the output pyramid with the desired number of levels and scale factor.
  uint32_t w, h;
  
  vpiImageGetSize(input, &w, &h);
  
  VPIImageType type;
  
  vpiImageGetType(input, &type);
  
  VPIPyramid output;
  
  vpiPyramidCreate(w, h, type, 4, 0.5, 0, &output);
Processing phase
1. Submit the algorithm to the stream, along with the input image and output pyramid.
  vpiSubmitGaussianPyramidGenerator(stream, input, output);
2. Optionally, wait until the processing is done.
  vpiStreamSync(stream);

Limitations and Constraints

Constraints for specific backends supersede the ones specified for all backends.

All Backends

Input image and output pyramid must have the same image type
Input image and pyramid's base level must have same dimensions
Only scale=0.5 is supported (i.e. only dyadic pyramids can be generated)
The following input image types are accepted:

PVA

Every pyramid level's dimension must be at least 65x65 big.
The following input image types are accepted:
- VPI_IMAGE_TYPE_U16
Pyramid must have at most 10 levels.

Performance

For further information on how performance was benchmarked, see Performance Measurement.

Jetson AGX Xavier
size	type	scale	levels	CPU	CUDA	PVA
1920x1080	u8	0.5	4	0.65 ms	0.098 ms	n/a
1920x1080	u16	0.5	4	0.74 ms	0.1130 ms	1.0756 ms

Jetson TX2
size	type	scale	levels	CPU	CUDA	PVA
1920x1080	u8	0.5	4	1.8 ms	0.256 ms	n/a
1920x1080	u16	0.5	4	2.3 ms	0.255 ms	n/a

Jetson Nano
size	type	scale	levels	CPU	CUDA	PVA
1920x1080	u8	0.5	4	3.441 ms	0.3375 ms	n/a
1920x1080	u16	0.5	4	3.710 ms	0.463 ms	n/a

VPI - Vision Programming Interface