Introducing FastCV: computer vision technology, tuned for mobile

Tuesday 10/25/11 12:00pm
|
Posted By Sy Choudhury
  • Up0
  • Down0

We love the smarts that come with developers adding computer vision-enabled functionality into their cutting edge applications, but we needed the processing to happen faster. That camera on your mobile phone pulls in a lot of data, so we asked ourselves how we could retain the smarts while still processing the information efficiently so that users get great experiences without bogging down the device. That's what led Qualcomm to launch FastCV, the first mobile-optimized computer vision (CV) library for ARM-based devices. Qualcomm is making it available to the industry so that CV applications can run efficiently on a wide array of mobile devices, even mass market handsets.

FastCV offers the most widely used, computationally intensive vision processing APIs, with hardware acceleration and better performance on mobile devices. FastCV is designed for efficiency on all ARM-based processors, but is also especially tuned to take advantage of Qualcomm's Snapdragon™ family of processors (second generation and above). In fact, FastCV for Snapdragon is able to offload processing onto various subsystems within the processor, whereas FastCV for ARM is limited to acceleration on the CPU.

FastCV is designed for developers interested in creating sophisticated CV apps, as well as CV middleware developers looking to build the frameworks necessary for everyday developers to include computer vision functionality in their apps.

Of course, FastCV is the framework at the heart of our vision-based Augmented Reality (AR) SDK, because AR is much more precise and useful when it's based on camera input than on location-based estimates. We anticipate FastCV will be used by middleware developers to build additional frameworks that will allow developers of computer vision apps to build compelling experiences around CV that take advantage of:

  • gesture recognition
  • face detection, tracking and recognition
  • text recognition and tracking
  • depth of field calculations

Here's what Daqri, a developer who was given early access to the FastCV API, has to say about it:

"Qualcomm has really thought through the challenges and constraints of mobile devices to provide the industry with a smart, optimized solution for delivering rich CV functionality," said Brian Mullins, CEO for Daqri. "The FastCV tools will easily tackle some of the harder aspects of computer vision technology, allowing us to focus on building compelling marketing platforms."

We're releasing the initial FastCV SDK for Android at the IEEE ISMAR conference in Basel, Switzerland Oct 26 - 29, but you don't have to fly there to get it. The FastCV SDK is available for download now at http://developer.qualcomm.com/fastcv.

Stay tuned for versions on other platforms, and let us know what you think in the forums, or in comments below.