Deep Learning Platform for High Performance Machine Learning for Automatic Target Detection and Recognition
Abstract
Major Goals: This project funded the acquisition of a Deep Learning Platform for Big Data (DLP-BD) with high performance data storage, networking, and processing capabilities to enable machine learning and specifically Deep Learning for ATR. The acquisition responds to the need for an in-house instrument enabling permanent storage of sensitive data that are in the tens of terabyte-range, and providing efficient big data processing through a non-volatile memory enabled hierarchical storage system, high speed networking interconnects, and multiple processors in parallel. Accomplishments: During the past year, the acquired Deep Learning Platform for Big Data (DLP-BD) has proved crucial to our DoD project that involved developing algorithms for the detection, recognition, and tracking of targets (vehicles and dismounts) using InfraRed (IR) imagery. The main objective of this project is to support the US Army Advanced Targeting and Lethality Automated System , ATLAS, program. ATLAS is a joint program, between the US Army C5ISR center and the Armament center. It aims to develop autonomous target acquisition technology, that will be integrated with fire control technology, aimed at providing ground combat vehicles with the capability to acquire, identify, and engage targets at least 3X faster than the current manual process. Our work in this area is in collaboration with the Army's Night Vision and Electronic Sensors Directorate, and more specifically the ATLAS team. The acquired DLP-BD have enabled local storage of large ITAR data and the adaptation, training, optimization, and analysis of various Deep Network architectures. Consequently, we have developed robust target detection and identification algorithms that continue to have excellent performance in various blind testing experiments. The acquired DLP-BD is also currently being used to develop, train, and analyze a deep learning algorithm for door detection in an indoor setting.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 31, 2021
- Accession Number
- AD1199850
Entities
People
- Hichem Frigui
Organizations
- University of Louisville