Binary Hashing with Semidefinite Relaxation and Augmented Lagrangian

This paper proposes two approaches for inferencing binary codes in two-step (supervised, unsupervised) hashing. We first introduce an unified formulation for both supervised and unsupervised hashing.