TOOLS AND PROCESSORS FOR COMPUTER VISION Selected Results from the Embedded Vision Alliance s Spring 2017 Computer Vision Developer Survey 1
EXECUTIVE SUMMARY Since 2015, the Embedded Vision Alliance has surveyed computer vision developers regarding the products they are working on and the hardware and software tools they are using in their projects. This white paper provides selected results from our most recent survey, conducted in. We received responses from 668 computer vision developers across a wide range of industries, organizations, geographical locations, and job types. We have focused our analysis on the 238 respondents whose organizations are developing end products for consumers, businesses, or governments (vs. organizations that are providing services, or providing components, subsystems, or software for incorporation into new products). We hope these selected results provide insight into the popular hardware and software platforms being used today for computer vision end-product creation. Full survey results are available to Embedded Vision Alliance Member companies. The Embedded Vision Alliance is a worldwide industry partnership bringing together technology providers and endproduct companies who are enabling innovative and practical applications for computer vision. The mission of the Alliance is two-fold: 1. Inspire and empower product creators to incorporate visual intelligence into new products and applications, and 2. Enable Alliance member companies to accelerate success in computer vision. For information on joining the Embedded Vision Alliance, please visit www.embedded-vision.com Note: Percentages add up to more than 100% in many of the charts presented in this white paper. This is because many of the questions asked respondents to select more than one option. 2017 Embedded Vision Alliance. All rights reserved. 1
CURRENTLY USING OR PLANNING TO USE COMPUTER VISION IN YOUR PRODUCTS OR SERVICES 6% 5% Not planning to use Don't know 19% Planning to use 70% Currently using Almost 90% of survey respondents are developing or planning to develop computer vision products. Similar to our 2016 survey, the overwhelming majority of respondents are developing or planning to develop products using computer vision. 2
TYPES OF PROCESSORS USED FOR VISION TASKS Ranked as One of Top Three 80 70 60 50 40 30 20 10 0 * Not asked in 2016 72% 67% -3-3 CPU GPU 40% +9 FPGA FPGAs are gaining in popularity. Despite programming challenges, they can deliver extremely high performance on demanding computer vision algorithms. 37% -7 Mobile 29% -2 Visionspecific 21% 19% -5 DSP Dedicated deeplearning processor* Note: This year we added a new choice ( deep-learning processors ) to this question. Since this question asks respondents to rank order the processors they are using, adding a choice tends to cause a decrease across all categories (hence the downward arrows across the board.) 3
PROGRAMMING LANGUAGES USED FOR VISION TASKS Ranked as One of Top Three 80 77% C++ remains the most popular language for computer vision development. 70 60 50 40 30 +1 46% +17 42% 40% -5 +7 37% But Python is rapidly gaining, having moved C out of second place since 2016. 29% 20 10 +7 9% 11% +2 C++ Python C OpenCL MATLAB CUDA Objective-C Other CUDA and OpenCL are gaining in popularity. Both CUDA and OpenCL registered gains, suggesting increased use of GPUs for computer vision. 4
LIBRARIES AND APIs USED FOR VISION DEVELOPMENT Ranked as One of Top Three 100 80 89% +7 OpenCV dominates development. An astonishing 89% of developers report it as one of their top-three computer vision libraries or APIs. OpenCV 3.0 adds functionality (e.g., deep learning), which is likely responsible for some of its increase in popularity. 60 43% 40 20-4 25% -1 18% -8 12% -3 OpenCV OpenGL OpenVX FastCV Other 5
USE OF NEURAL NETWORKS FOR COMPUTER VISION 17% Yes, extensively 16% Yes, in a minor role 4% Don t know 17% No 40% Not yet, but planning to Almost 80% of survey respondents use or plan to use neural networks. In the last several years neural networks have emerged as one of the dominant techniques for computer vision. 6
DESIGN SOFTWARE FOR NEURAL NETWORKS Ranked as One of Top Three 70 60 50 64% 61% +37 +14 TensorFlow steals the show. Since its introduction in late 2015, Google s open-source TensorFlow has emerged as the most popular deep learning/neural network design framework, displacing all others. 40 30 41% 36% +26-12 20 10 15% 14% 12% +3 +2-3 0 TensorFlow Caffe OpenCV MATLAB Torch Theano Other Depsite this, Caffe and OpenCV remain strong. Many developers still use Caffe for neural network design, and OpenCV s deep learning module is also popular. 7
OTHER ALLIANCE PROGRAMS VISION ACCELERATOR PROGRAM The Vision Accelerator Program helps companies quickly understand and navigate the technical and business complexity of incorporating visual perception capabilities so they can more quickly and confidently plan, develop and deliver their products. It is a service available to members of the Embedded Vision Alliance who are developing end products and systems with visual perception capabilities (e.g., deep learning, 3D sensing). The Vision Accelerator Program helps companies: Make decisions in a fast-changing market where areas like deep learning and 3D sensing are rapidly moving from research into practical use Understand the tradeoffs for low-power, lowcost devices and cloud processing Know what vision software standards, open source tools and algorithms are gaining traction Identify which startups, suppliers, partners and experts have relevant vision technologies and know-how Build skills and recruit the right talent Access and develop a network of experts, suppliers and partners For more information on the Vision Accelerator Program, please email accelerate@embedded-vision.com EMBEDDED VISION SUMMIT The Embedded Vision Summit, held in Silicon Valley every year in May, is the only event focused exclusively on the technologies, hardware, and software that bring visual intelligence to products. The Summit presents the latest practical techniques and technologies for vision-based product development, and illuminates the commercial landscape, trends, and business opportunities in this fast-growing market. It inspires participants to use vision technology in new ways and empowers them with the knowhow they need to integrate vision capabilities into products. The 2017 event featured more than 90 expert presenters in 4 conference tracks covering every aspect of computer vision. The event s Vision Technology Showcase included more than 100 demonstrations of commercially-available computer vision components and solutions both in hardware and software from more than 50 top suppliers. Day 3 of the 2017 conference consisted of in-depth hands-on Vision Technology Workshops presented by Alliance Member companies. For more information on the Embedded Vision Summit, please visit www.embedded-vision.com/summit
1646 North California Blvd Suite 220 Walnut Creek, CA 94596 USA Phone: +1 (925) 954-1411 Fax: +1 (925) 954-1423 info@embedded-vision.com 9