Convolutional Neural Networks is renowned for its state-of-the-art capability in intelligently filtering, differentiating, and recognizing visual information (i.e. shapes, colors). Recent developments in convolutional neural networks approaches using FPGAs have significantly boosted the performance of image processing/recognition systems. See how these excellent designs accelerate these networks with the powerful Intel Cyclone V SoC FPGA!

An unprecedented FPGA design festival co-organized by Terasic, Intel®, Intel® FPGA University Program, DigiKey, and sponsored by Analog Devices and ISSI, 2018 InnovateFPGA Global Design Contest has now reached the last week of semi-final judging period! Beginning today, we’ll pick out exciting designs weekly to introduce to global FPGA communities. Join the community now to watch the teams delivering their one-of-a-kind innovation! Being a part of the community, you can interact with teams by offering advice, encouragement & starting a conversation. Active community members will get a chance to win the most coveted Terasic kit, the DE10-Nano,－a kit that just may change your FPGA development career forever!

2018 InnovateFPGA Grand Final:

2018 InnovateFPGA Grand Final Grand Final will take place in Intel Campus, San Jose, USA from Aug 14~ Aug 15. We welcome all developers and FPGA enthusiasts to attend the event! More details of Grand Final will be released in June. Stay tuned!

AS031» BreXting : Brain Texting
Authors: McGill University, Ecole Polytechnique de Montreal

This system is designed to record brain activity, predict the characters pop up in mind and display on screen in real time. The prediction is based on machine learning using EEG. The system has 14 sensors to record the brain wave. The EEG signal is transmitted through bluetooth and recveived by DE10-Nano via Arduino interface. The CNN is implemented on DE10-Nano for decoding.

The system is meant to connect human brain with host PC and is useful for those who have been suffering brain damage.

AS037» CRIOS - Cruise speed controller with Deep Learning
Authors: Ânderson Ignacio da Silva

The design targets a comprehensive system to inspect traffic signs through deep learing and fast-forward split image processing. The categorized result will be informed to driver for maintaining the course and speed control. The system covers three key parts: identify traffic signs in real time, categorize traffic signs based on CNN implemented on FPGA, and report back to driver for assistance.

The system can be the fundamental element from self-learning automotive to the automatic identification of traffic signs.

PR061» Diagnostic System based on Deep Learning for Eye Disease
Authors:Sun Yat-sen University

The system collects eye diagrams from digital camera and uses FPGA for the hardware acceleration. The diagnose result is calculated on DE10-Nano using CNN.

This system can effectively assist doctors and increase the efficiency of diagnostic to obtain correct result in real time through artificial intelligence.

PR086» Intelligent Control System based on the Analysis of Target Behavior
Authors: Changchun University of Technology

This system collects, processes, save, and read video and voice signals using FPGA. The data is exchanged between FPGA and ARM to complete a series of analysis based on target behavior such as emotion identification and body language.

The system can be broadly used in social and personal security under the scenarios such as home and business.

PR051»BioVision
Authors: National Taiwan University of Science and Technology

The design captures facial images via 8 megapixel digital camera and processes the data using machine learning algorithm to work out the parameters of vital signs. The result is displayed on a HDMI monitor and it can be transmitted to Android smartphone through Bluetooth for further visualization and coding.

The system adopts visual technology via high specification yet low cost and it can be further developed for robotic, medical machinery, and intelligence assistance technology.

AP074» Detection and Recognition of Plant Diseases using FPGA based real-time processing
Authors: University of Moratuwa

This system captures images of plants through USB camera and saves the data to SDRAM for HPS to run pre-processing and analysis. The processed information is sent to the CNN running on FPGA and displayed on a HDMI touchscreen.

The purpose of this design intends to detect the symptoms of sick plants by applying image processing technology. This can benefit farmers and researchers who are specialized in plant disease.

EM112 » Intelligence Guiding System based on Machine Learning
Authors: Addis Ababa University

This system captures images through digital camera and run the machine learning algorithm on FPGA to locate the obstacles in the front. The signals will be transmitted to the wrist band of passenger via Bluetooth and guide blind person to the safe path via different level of vibration.

The system can be implemented on portable device in large scale to help people with limited visibility and guide a safer way home.