Combination of tangent vectors and local representations for handwritten digit recognition

Abstract

Statistical classification using tangent vectors and classification based on local features are two successful methods for various image recognition problems. These two approaches tolerate global and local transformations of the images, respectively. Tangent vectors can be used to obtain global invariance with respect to small affine transformations and line thickness, for example. On the other hand, a classifier based on local representations admits the distortion of parts of the image. From these properties, a combination of the two approaches seems veryl ikely to improve on the results of the individual approaches. In this paper, we show the benefits of this combination byap plying it to the well known USPS handwritten digits recognition task. An error rate of 2.0% is obtained, which is the best result published so far for this dataset.