The method proposed offers a significant advantage in computation time for an image classification system. It may be pssible in addition to reduce information storage requirements. An important consideration in today's tactical environment is to be able to classify quickly an object in a scene. This method encourages the examination of an image feature data beforehand in the lab to understand feature and parameter realtionships. This permits efficient operation in a field environment. The older approaches blindly accumulate and store data in the lab, and wait until the system is in the field to look through it. Clearly, the proposed algorithm should be considered by future systems designers.
Eliminating nearest neighbor searches in estimating target orientation
Schaetzung der Zielrichtung durch Aussonderung des naechsten Nachbarn
1982
10 Seiten, 18 Bilder, 2 Tabellen, 2 Quellen
Conference paper
English
British Library Online Contents | 2015
|Nearest neighbor search for relevance feedback
IEEE | 2003
|LDA/SVM driven nearest neighbor classification
IEEE | 2001
|Nearest Neighbor Search for Relevance Feedback
British Library Conference Proceedings | 2003
|