Adaptive Computing Developer Contest

A new flexible hardware architecture to enable adaptive support for a variety of DL based video analytics algorithms in extreme environments

We have converted, trained, tested, quantized, compiled and run enet, densenet and yolo models on ultra96 successfully.

Social distance monitor using yolov3 and DPU acceleration on Ultra96v2

Xilinx AI-based Otitis media detection to assist clinicians in ear disease diagnosis using Deep Learning-based annotated Image Retrieval.

Recognize what material an object is made of by simply touching it

FPGA implementation of a highly efficient real-time machine learning driving assistance application using a camera circuit.

This project will demonstrate how to perform object detection using a Ultra96v2 board linked to a Vector robot. The used algorithm is YOLOv3

Build real-time face mask detection application running on Ultra96-V2 board using Vitis-AI.

We aim to use HW acceleration and the power of AI to detect COVID-19 infection in a jiffy using X-rays.

Improvising on Precision Farming technique, through combined efforts of DL & sensing Electrophysiological response from plant.

Humans could worry less about finding parking spots if there were a smart overhead vision system in every parking lot to locate the spots.

AI on the Ultra96V2 board, for inspiring kids math study interest, playing math game such as 24 points.

An object detection model running on the Ultra96v2 providing haptic feedback for independent mobility of the visually challenged

Use ZCU104's video analysis capabilities to allow businesses and schools to track whether masks are being worn in public places.

Develop Intelligent Video Processing Applications on Zynq UltraScale+ devices in minutes.

Our goal is to make medical data storage systems more efficient by introducing high compression ratio lossless image compression.

This project uses XIlinx ZCU104 to monitor fall detection using Vitis-AI and sending Alarm via SIren and IFTTT.

Based on our SPO2 levels, the R and G channels of our skin show variations while breathing that can be used to estimate the SPO2 levels.

Real-time vehicle distance estimation is an extremely challenging problem. DriverPal V2 solve this issue by implementing a two stage system.

FPGA-based system that monitors facemask use through artificial intelligence, includes a thermometer and facemask dispenser.

Inference at the edge is the new hype. But what about training at the edge ? This project aims at exploring that with FPGA acceleration.

Quantifying uncertainty in autonomous driving may safe lifes. If we do this inference fast, it gets even better.

In this project we deploy both online and offline TSM networks to FPGA for video understanding tasks through 2D CNN.

The power consumption is only 2.3 W, and the accuracy rate can reach 95%, which is better than the existing circuits of the same scale.