Eric Gregori, Senior Software Engineer at BDTI, presents the "Accelerating Computer Vision Applications with the Hexagon DSP" tutorial at the September 2013 Qualcomm UPLINQ Conference.
Smartphones, tablets and embedded systems increasingly use sophisticated vision algorithms to deliver capabilities like augmented reality and gesture user interfaces. Since vision algorithms are computationally demanding, a key challenge when implementing vision in battery-powered devices is achieving energy-efficient processing. This BDTI tutorial presents a straightforward approach for using the Hexagon DSP core in Qualcomm’s Snapdragon application processors to offload vision processing functions from the CPU, in order to reduce power and free up CPU cycles for other tasks.
