SIFT Implementation and Optimization for General-Purpose GPU

Abstract

With the addition of free programmable components to modern graphics hardware, graphics processing units
\n(GPUs) become increasingly interesting for general purpose computations, especially due to utilizing parallel
\nbuffer processing. In this paper we present methods and techniques that take advantage of modern graphics
\nhardware for real-time tracking and recognition of feature-points. The focus lies on the generation of feature
\nvectors from input images in the various stages. For the generation of feature-vectors the Scale Invariant Feature
\nTransform (SIFT) method [Low04a] is used due to its high stability against rotation, scale and lighting condition
\nchanges of the processed images. We present results of the various stages for feature vector generation of our
\nGPU implementation and compare it to the CPU version of the SIFT algorithm. The approach works well on
\nGeforce6 series graphics board and above and takes advantage of new hardware features, e.g. dynamic
\nbranching and multiple render targets (MRT) in the fragment processor [KF05]. With the presented methods
\nfeature-tracking with real time frame rates can be achieved on the GPU and meanwhile the CPU can be used for
\nother tasks.