CVMay 22, 2014

Semi-supervised Spectral Clustering for Classification

arXiv:1405.5737v2
Originality Incremental advance
AI Analysis

This work addresses image-set classification, an incremental improvement for computer vision applications.

The authors tackled the problem of image-set classification by proposing a Classification Via Clustering (CVC) algorithm that co-clusters training and test data to assign labels, and it outperformed state-of-the-art methods on five standard datasets.

We propose a Classification Via Clustering (CVC) algorithm which enables existing clustering methods to be efficiently employed in classification problems. In CVC, training and test data are co-clustered and class-cluster distributions are used to find the label of the test data. To determine an efficient number of clusters, a Semi-supervised Hierarchical Clustering (SHC) algorithm is proposed. Clusters are obtained by hierarchically applying two-way NCut by using signs of the Fiedler vector of the normalized graph Laplacian. To this end, a Direct Fiedler Vector Computation algorithm is proposed. The graph cut is based on the data structure and does not consider labels. Labels are used only to define the stopping criterion for graph cut. We propose clustering to be performed on the Grassmannian manifolds facilitating the formation of spectral ensembles. The proposed algorithm outperformed state-of-the-art image-set classification algorithms on five standard datasets.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes