For rapid development of high performance image processing applications, Aspex provides a parameterizable image processing library, callable as C functions.
1D filters
-
FilterM - One dimensional FIR/convolution filter
-
BoxM - One dimensional box filter
-
Gauss3 - one dimensional 3-tap Gaussian filter
-
GaussM - one dimensional M-tap Gaussian filter
|
2D kernel filters
 |
|
|
-
FilterMxN - Two dimensional FIR filter
-
BoxMxN - Two dimensional box filter
-
GaussMxN - M x N Gaussian filter
-
Gauss3x3 - 3 x 3 Gaussian filter
-
Sobel - Sobel filter
-
Laplace - Laplacian filter
-
Median3x3 - 3x3 median filter
|
Other functions
 |
- 256 / 512 / 1024 / 2048 / 4096 / 8192 point
|
- 2D FFT / iFFT
- 2D Wavelet transform
- Discrete Cosine Transform / iDCT
- Colour space conversion
|
Custom library functions
Our team of application engineers can quickly generate new library functions to perform specific tasks. If you have an application requirement which is not listed above, please contact us.
Parameterisable size and precision
For most of the functions in the Aspex Image Processing Library, the size of the region of interest and the arithmetic precision to be used are defined by C variables or constants, allowing compile-time or run-time configuration of the functions.
For example, the parameters of the generic M x N 2D kernel convolution filter are as follows:
- Input, Output - vector variable type, defined by start address and precision
- int KernelWidth - width of kernel filter (pixels)
- int KernelHeight - height of kernel filter (pixels)
- int Kernel[KernelHeight][KernelWidth] - C array of filter coefficients
|
|
|
