Texture Analysis Using Multidimensional Histogram

Texture features have long been used in remote sensing applications for representing and retrieving regions similar to a query region. Various representations of texture have been proposed based on the power spectrum, grey-level cooccurrence matrices, wavelet features, Gabor features, etc. Analysis of several co-occurring pixel values may benefit texture description but is impeded by the exponential growth of histogram size. Multidimensional histograms can be reduced by using methods like linear compression, dimension optimization and vector quantization. Experiments with natural textures showed that multidimensional histograms provided higher classification accuracies than the channel histograms and the wavelet packet signatures.