A Method of Airplane Recognition Based on Supervised Locally Linear Embedding

Dimensionality reduction is an important preprocessing step before object recognition.Since the airplane objects are very sensitive to the various geographic transformations(such as translation,rotation,affine transform,etc) and the influence of observed angle,locality and illumination in the real scene,many linear dimensional reduction methods are not effective to recognize the airplane objects.The Locally Linear Embedding(LLE) algorithm is a useful and effective non-linear dimensional reduction algorithm.In this paper,a Supervised LLE(SLLE) was proposed for the complex airplane object recognition.The experiment results showed that the proposed method is very effective and feasible.