Mission Driven Scene Understanding: Candidate Model Training and Validation

Abstract

Army missions take place in dynamic environments, where changing illumination, precipitation, and vegetation can modify saliency and context of an outdoor scene, obscure features, and degrade object recognition. For Army missions, scene understanding tools need to account for dynamic environments that change as a function of space and time and should be tested in mission simulating conditions. In addition, the impact of dynamic environments should be included in the scene understanding approach. At this stage, we are evaluating different computational frameworks that may be useful to incorporate dynamic environments into mission driven scene understanding. One of the candidate engines that we are evaluating is a convolutional neural network (CNN) program installed on a Windows 10 notebook computer. In this report, we present progress toward the proof-of-principle testing of the candidate model to examine the impact of dynamic environments on scene understanding model results.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2016
Accession Number
AD1016729

Entities

People

  • Arnold D. Tunick

Organizations

  • United States Army Research Laboratory

Tags

Communities of Interest

  • Autonomy
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Artificial Intelligence Software
  • Central Processing Units
  • Computer Languages
  • Computer Programming
  • Computer Programs
  • Computer Vision
  • Computers
  • Convolutional Neural Networks
  • Data Sets
  • Deep Learning
  • Information Science
  • Laptop Computers
  • Neural Networks
  • Object Recognition
  • Recognition
  • Training
  • Validation

Fields of Study

  • Computer science

Readers

  • Computer Vision.
  • Military Science and Technology Research and Modernization.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Space