Accelerating 3D convolution using graphics hardware (case study)

Many volume filtering operations used for image enhancement, data processing or feature detection can be written in terms of three-dimensional convolutions. It is not possible to yield interactive frame rates on todays hardware when applying such convolutions on volume data using software filter routines. As modern graphics workstations have the ability to render two-dimensional convoluted images to the frame buffer, this feature can be used to accelerate the process significantly. This way generic 3D convolution can be added as a powerful tool in interactive volume visualization toolkits.