Yurii Piadyk, Ph.D.

Yurii Piadyk, Ph.D.

Research Associate at New York University

Publications

StreetAware: A High-Resolution Synchronized Multimodal Urban Scene Dataset

Published in MDPI Sensors, 2023

Data collected from sensors in areas of high activity in the urban environment is valuable for researchers interpreting the dynamics between vehicles, pedestrians, and the built environment. We present a high-resolution audio, video, and LiDAR dataset of three urban intersections in Brooklyn, New York, totaling almost 8 unique hours. The data was collected with custom REIP sensors that were designed with the ability to accurately synchronize multiple video and audio streams.

Download here

REIP: A Reconfigurable Environmental Intelligence Platform and Software Framework for Fast Sensor Network Prototyping

Published in MDPI Sensors, 2022

Sensor networks have dynamically expanded our ability to monitor and study the world, and their presence and need keep increasing. We introduce REIP, a Reconfigurable Environmental Intelligence Platform for fast sensor network prototyping. REIP’s first and most central tool, implemented in this work, is an open-source software framework with a flexible modular API for data collection and analysis using multiple sensing modalities.

Download here

Hardware Design and Accurate Simulation of Structured-Light Scanning for Benchmarking of 3D Reconstruction Algorithms

Published in NeurIPS 2021 Datasets and Benchmarks Track (Round 2), 2021

Images of a real scene taken with a camera commonly differ from synthetic images of a virtual replica of the same scene, despite advances in light transport simulation and calibration. By explicitly co-developing the Structured-Light Scanning (SLS) hardware and rendering pipeline we are able to achieve negligible per-pixel difference between the real image and the synthesized image on geometrically complex calibration objects with known material properties.

Download here

Tracking Sparse Objects and People in Large Scale Environments

Published in United States Patent and Trademark Office, 2021

A modular tracking system is described comprising of the network of independent tracking units. Markerless computer vision algorithms are executed directly on the units and provide feedback to motorized mirror placed in front of the zoomed camera to keep tracked objects/people in its field of view. Inference from different sensor is fused in real time to reconstruct high-level events and full skeleton representation for each participant.

Download here

LegoTracker: An Intelligent Modular System for Large-Scale Sports Tracking

Published in NVIDIA GPU Technology Conference, 2020

For past decades sports tracking was limited to a rough representation of each player by a single point and often relies on special markers integrated into sports apparel. We propose a novel modular sports tracking system comprising of independent units, each running state-of-the-art algorithms for player detection and tracking, and which provides a full skeleton representation for each player over a large game field.

Download here

Anisotropic Subsurface Scattering Acquisition Through a Light Field Based Apparatus

Published in International Symposium on Electronic Imaging: Imaging Sensors and Systems, 2020

Subsurface scattering gives a distinct look to many everyday objects. We demonstrate a system that can quickly acquire the full anisotropic subsurface scattering for homogeneous materials. Unlike many existing commercial acquisition systems, our system can be assembled from off-the-shelf optical component and 3D printed/cut parts.

Download here

A pattern recognition mezzanine based on Associative Memory and FPGA technology for L1 track triggering at HL-LHC

Published in Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2016

We present the tests of a prototype system (Pattern Recognition Mezzanine) as core of pattern recognition and track fitting for HL-LHC ATLAS and CMS experiments, combining the power of both Associative Memory custom ASIC and modern Field Programmable Gate Array (FPGA) devices.

Download here

FTK AMchip05: an Associative Memory Chip Prototype for Track Reconstruction at Hadron Collider Experiments

Published in European Physical Society Conference on High Energy Physics, 2015

This presentation is about ATLAS FTK track trigger project based on FPGAs and a custom ASIC: the Associative Memory chip (AMchip). The AMchip is a core processor in charge of the real-time pattern recognition stage of the FTK algorithm. It includes Content Addressable Memory with advanced computation logic enabling detection of correlated data patterns (e.g. tracks made of hits) within a sparse dataset at full I/O speed.

Download here