&lt;title&gt;AdaptSAPS utility for adaptive ATR development and assessment&lt;/title&gt;

Williams, Arnold C.; Pachowicz, Peter W.; Leonard, Jim

doi:10.1117/12.555515

Cited by 2 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These compare favourably with the results from[8], which give average mission results for Mission 6 of D = 0.577, E = 0.224, and MOP-A = 0.324 (the Pcc metric is not given).…”

supporting

confidence: 74%

“…In this way, it is possible to examine the degree of improvement and adaptation that the system undergoes over the course of a mission. The mission results at [8] provide an indication of a baseline level of performance for a simple system that discriminates between target and non-target vehicles. For 'Mission 6', the Discrimination metric D was 0.577, the reported accuracy E was 0.224, and the MOP-A metric was 0.324.…”

Section: Previous Adaptsaps Systemsmentioning

confidence: 99%

“…Whilst AdaptSAPS provides pre-defined baseline missions and well defined metrics for this very reason, only one of the missions (mission 6) presented at [8] contains target and non-target vehicles; the rest only contain target vehicles and clutter. The metrics are described in [4], which defines a set of performance measures for a system's adaptability.…”

Section: Previous Adaptsaps Systemsmentioning

confidence: 99%

“…However, research on operator feedback using the AdaptSAPS framework has not been widely published to date. [8] has used the AdaptSAPS system to develop a simple adaptive ATR system in order to evaluate the performance of AdaptSAPS, and determine what changes might be required for future versions of the program. This system only attempts to separate T-72 Tanks from clutter and confusers, and does not use the rest of the target vehicles.…”

Section: Previous Adaptsaps Systemsmentioning

confidence: 99%

See 3 more Smart Citations

Utilizing feedback in adaptive SAR ATR systems

Horsfield¹,

Blacknell

2009

SPIE Proceedings

View full text Add to dashboard Cite

Existing SAR ATR systems are usually trained off-line with samples of target imagery or CAD models, prior to conducting a mission. If the training data is not representative of mission conditions, then poor performance may result. In addition, it is difficult to acquire suitable training data for the many target types of interest. The Adaptive SAR ATR Problem Set (AdaptSAPS) program provides a MATLAB framework and image database for developing systems that adapt to mission conditions, meaning less reliance on accurate training data. A key function of an adaptive system is the ability to utilise truth feedback to improve performance, and it is this feature which AdaptSAPS is intended to exploit. This paper presents a new method for SAR ATR that does not use training data, based on supervised learning. This is achieved by using feature-based classification, and several new shadow features have been developed for this purpose. These features allow discrimination of vehicles from clutter, and classification of vehicles into two classes: targets, comprising military combat types, and non-targets, comprising bulldozers and trucks. The performance of the system is assessed using three baseline missions provided with AdaptSAPS, as well as three additional missions. All performance metrics indicate a distinct learning trend over the course of a mission, with most third and fourth quartile performance levels exceeding 85% correct classification. It has been demonstrated that these performance levels can be maintained even when truth feedback rates are reduced by up to 55% over the course of a mission.

show abstract

“…These compare favourably with the results from[8], which give average mission results for Mission 6 of D = 0.577, E = 0.224, and MOP-A = 0.324 (the Pcc metric is not given).…”

supporting

confidence: 74%

Section: Previous Adaptsaps Systemsmentioning

confidence: 99%