A fast and efficient sift detector using the mobile gpu

Emerging mobile applications, such as augmented reality, de-mand robust feature detection at high frame rates. We present an implementation of the popular Scale-Invariant Feature Transform (SIFT) feature detection algorithm that incorpo-rates the powerful graphics processing unit (GPU) in mobile devices. Where the usual GPU methods are inefficient on mo-bile hardware, we propose a heterogeneous dataflow scheme. By methodically partitioning the computation, compressing the data for memory transfers, and taking into account the unique challenges that arise out of the mobile GPU, we are able to achieve a speedup of 4-7x over an optimized CPU version, and a 6.4x speedup over a published GPU imple-mentation. Additionally, we reduce energy consumption by 87 percent per image. We achieve near-realtime detection without compromising the original algorithm. Index Terms — computer vision, mobile computing, fea-ture detection, graphics processing unit (GPU), OpenGL for Embedded Systems (OpenGL ES) 1