KEYWORDS: Visualization, Palladium, Enhanced vision, Human vision and color perception, Synthetic vision, Visual process modeling, Nonlinear filtering, Sensors, Target detection, Color vision
One of the key problems in developing Enhanced and Synthetic Vision Systems is evaluating their effectiveness in enhancing human visual performance. A validated simulation of human vision would provide a means of avoiding costly and time-consuming testing of human observers. We describe an image-based simulation of human visual search, detection, and identification, and efforts to further validate
and refine this simulation. One of the advantages of an image-based simulation is that it can predict performance for exactly the same visual stimuli seen by human operators. This makes it possible to assess aspects of the imagery, such as particular types and amounts of background clutter and sensor distortions, that are not usually considered in non-image based models. We present two validation studies - one showing that the simulation accurately predicts human color discrimination, and a second showing that it produces probabilities of detection (Pd's) that closely match Blackwell-type human threshold data.
Biologically-based computer vision systems are now available that achieve robust image interpretation and automatic target recognition (ATR) performance. We describe two such systems and the reasons behind their robust performance. We also report results of three studies that demonstrate this robustness.
KEYWORDS: Visual process modeling, Machine vision, Algorithm development, Inspection, Visualization, Image processing, Spatial frequencies, Systems modeling, Process control, Visual system
The design of robust machine vision algorithms is one of the most difficult parts of developing and integrating automated systems. Historically, most of the techniques have been developed using ad hoc methodologies. This problem is more severe in the area of natural/biological products. In this arena, it has been difficult to capture and model the natural variability to be expected in the products. This present difficulty in performing quality and process control in the meat, fruit and vegetable industries. While some systems have been introduced, they do not adequately address the wide range of needs. This paper will propose an algorithm development technique that utilizes modes of the human visual system. It will address that subset of problems that humans perform well, but have proven difficult to automate with the standard machine vision techniques. The basis of the technique evaluation will be the Georgia Tech Vision model. This approach demonstrates a high level of accuracy in its ability to solve difficult problems. This paper will present the approach, the result, and possibilities for implementation.
The Georgia Tech Research Institute has developed an integrated suite of software for Visual and Electro-Optical (VISEO) detection analysis, under the sponsorship of the Army Aviation and Troop Command, Aviation Applied Technology Directorate. The VISEO system is a comprehensive workstation-based tool for multi-spectral signature analysis, LO design, and visualization of targets moving through real measured backgrounds. A key component of the VISEO system is a simulation of real measured backgrounds. A key component of the VISEO system is a simulation of human vision, called the Georgia Tech Vision (GTV) simulation. The algorithms used in the simulation are consistent with neurophysiological evidence concerning the functions of the human visual system, from dynamic light adaptation processes in the retinal receptors and ganglia to the processing of motion, color, and edge information in the striate cortex. The simulation accepts images seen by the naked eye or through direct-view optical systems, as well as images viewed on the displays of IR sensors, image intensifiers and night-vision devices. GTV outputs predicted probabilities that the target is fixated (Pfix) during visual search, and detected (Pd), and also identifies specific features of the target that contribute most to successful search and detection performance. This paper outlines the capabilities and structure of the VISEO system, emphasizing GTV. Example results of visible and IR signature reduction on the basis of VISEO will be shown and described.
A simulation of human pattern recognition is outlined which classifies objects based on outputs of a computational vision model, called the Georgia Tech Vision (GTV) model. It is shown that the simulation is able to identify high- level features of military targets, and that identification of high-level features can be used as a tool for recognizing targets. The results suggest that the computational vision model will simplify the task of simulating target recognition by providing a 'front-end' that simulates the basic features that human observes use to recognize targets.
Work in progress at Georgia Tech to develop a model of human pattern perception, visual search, and detection is reviewed. The model's algorithms are based on research on low-level visual processes. Recent advances in the field have led to the development of computational models of the image processing performed by the visual system from the cornea to the striate cortex. The model also incorporates recent advances from research on visual search. The organization of the model for predicting target acquisition, analyzing target signatures, and specifying low-observable requirements is discussed.
Previous sensor/observer performance prediction models have not explicitly treated false alarm effects, especially in relation to clutter. A method for predicting probabilities of detection as a function of observer false alarm probability and clutter is presented. The method is applied to previously collected observer data to determine the effect of observer false alarm probability on resolution criteria associated with specified clutter levels. The interdependent effects of clutter and observer false alarm probability are illustrated by deriving an expression for the probability of target acquisition in the case where the observer has only enough time to select one target candidate. It is concluded that false alarm probability has a significant impact on target resolution criteria in moderate and high background clutter. Example resolution criteria are given for several false alarm probabilities.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.