The author proposes the Latent Semantic Consensus (LSC) method for deterministic geometric model fitting, focusing on preserving latent semantic consensus in both data points and model hypotheses.
