Xing SUN


Large Scale Image Categorization in Sparse Nonparametric Bayesian Representation

A new hierarchical sparse coding algorithm is proposed, which combines hard and soft assignment coding in a fully unsupervised manner for image categorization. Hard coding assigns data into independent cluster globally to enable local dictionary learning to be learned by soft coding. The new coding algorithm is characterized by better fitting of data, more discriminative global clustering, low computational complexity and convergence speed up.

  • Hard coding

A local soft assignment coding for image category feature basis is updated in a sub-sparse coding optimization local iteration procedure.


  • Soft coding

In the image indicator optimization step, a hard assignment is applied to categorize different scenes. A global hard assignment coding for image category indicator is updated as following:


  • Sparse Nonparametric Bayesian model: