Uncanny Vision

Uncanny Vision develops optimized computer vision solutions based on embedded systems. Uncanny Vision provides two products UncannyCV and UncannyDL. UncannyCV is a computer vision library optimized for the Cortex-A series ARM processors, consisting of around 70 most commonly used computer vision/image processing functions. Uses a mixture of algorithm know-how and optimization skills and claims to achieve optimizations from 2x up to 20x, with average 5x optimization compared with reference software implementations. UncannyDL is an on-device deep learning library for mobile and Embedded applications. Supports any Convolutional Neural Network models based on the Caffe Model Zoo framework like AlexNet, GoogleNet, etc. These two products are available on a wide range of OS platforms for ARM Cortex-A series processors including Android, iOS, embedded Linux, various RTOS platforms. Selected by Microsoft Accelerator in September 2016, for its startup accelerator programme. Graduated from The Target Accelerator Program in September 2016.