Appearance Simulator for Computer Vision Research.

Abstract

Computer vision research needs a flexible and robust method to obtain accurate appearance of a scene. Such needs include generating test data, using a simulator as a part of an algorithm, and generating model data for object recognition and inspection. A sophisticated and reasonable mechanism for generating scene appearance is a technique known as ray tracing. However, current ray tracers are mainly developed for computer graphic research. They are based on ad hoc reflectance mechanisms; their algorithms emphasize computational costs by choosing computational expediency at the expense of an appearance grounded in physics. We are developing an appearance simulator to be able to produce a scene according to the laws of physics. By combining the reflectance theory based on physics and the technique of ray tracing, we are able to produce scene appearances that are adequate and realistic enough for computer vision. First, we will briefly describe our reflectance model to be used in our simulator. Then, we will discuss several implementation issues how to embody the model into real algorithms. Finally, we will generate several scene appearance to verify the performance of our system and illustrate applications of this system.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1991
Accession Number
ADA240506

Entities

People

  • Katsushi Ikeuchi
  • Shree Nayer
  • Yoshimasa Fujiwara

Organizations

  • Carnegie Mellon University

Tags

Communities of Interest

  • Air Platforms
  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Computer Vision
  • Computers
  • Flux Density
  • Geometry
  • Identification
  • Intensity
  • Light Sources
  • Object Recognition
  • Optics
  • Ray Tracing
  • Recognition
  • Reflectance
  • Simulations
  • Simulators
  • Specular Reflection
  • Surface Properties

Fields of Study

  • Computer science

Readers

  • Artificial Intelligence
  • Computational Modeling and Simulation
  • Image Processing and Computer Vision.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference