To achieve high classification accuracy of hyperspectral data, a dimensionality reduction algorithm called Sample-dependent repulsion graph regularized auto-enc
Noise reduction is a very important topic in image processing. We propose a new method to deal with the case where the noisy image has different noise levels in different regions. The main idea is to
In this paper, an hierarchical n-gram Language model(LM) combining words and characters is explored to improve the detection of Out-of-vocabulary(OOV) words in