Restricted Research - Award List, Note/Discussion Page

Fiscal Year: 2021

155  University of North Texas  (84451)

Principal Investigator: Yang,Qing

Total Amount of Contract, Award, or Gift (Annual before 2011): $ 99,948

Exceeds $250,000 (Is it flagged?): No

Start and End Dates: - 9/30/22

Restricted Research: YES

Academic Discipline: Computer Science & Engineering

Department, Center, School, or Institute: College of Engineering

Title of Contract, Award, or Gift: EAGER: SaTC: Privacy¿Preserving Convolutional Neural Network for Cooperative Perception in Vehicular Edge Systems

Name of Granting or Contracting Agency/Entity: National Science Foundation

Program Title: N/A
CFDA Linked: Computer and Information Science and Engineering


The overall research objective of this project is to design a privacy-preserving convolutional neural network on edge servers to efficiently process encrypted image data generated from vehicles in a vehicular edge system. Towards this end, two fundamental questions were identified that, once answered, achieve this objective. The approach to answering question 1 (“how to protect privacy when sensor data is processed and fused on edge servers in achieving cooperative perception”), is to design a privacy preserving convolutional neural network which extracts features from two ciphertexts to obtain the same object detection results as the original CNN. To address question 2 (“how to speed up secure object detection on edge servers while keeping the energy consumption low”), we will leverage cross-layer multiplication reduction and local- global addition reuse to optimize the privacy-preserving CNN, and customize an FPGA design to hardware program the optimized CNN model and accelerate the processing of encrypted images.

Discussion: No discussion notes


Close Window