__global__ void filter2D(int ofsX, int ofsY, PtrStepf dst, const int kWidth, const int kHeight, const int anchorX, const int anchorY, const BrdReflect101<float> brd)
template <class SrcT, typename D>
__global__ void filter2D(const SrcT src, DevMem2D_<D> dst, const int kWidth, const int kHeight, const int anchorX, const int anchorY)
void filter2D_gpu(DevMem2Df src, int ofsX, int ofsY, DevMem2Df dst, int kWidth, int kHeight, int anchorX, int anchorY, float* kernel, cudaStream_t stream)
template <typename T, typename D>
void filter2D_gpu(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst,
int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel,
int borderMode, const float* borderValue, cudaStream_t stream)
typedef void (*func_t)(const DevMem2D_<T> srcWhole, int xoff, int yoff, DevMem2D_<D> dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<uchar, uchar>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<uchar4, uchar4>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<ushort, ushort>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<ushort4, ushort4>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<float, float>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);
template void filter2D_gpu<float4, float4>(DevMem2Db srcWhole, int ofsX, int ofsY, DevMem2Db dst, int kWidth, int kHeight, int anchorX, int anchorY, const float* kernel, int borderMode, const float* borderValue, cudaStream_t stream);