Bayesian Object Localisation in Images

摘要

A Bayesian approach to intensity-based object localisation is presented that employs a learned probabilistic model of image filter-bank output, applied via Monte Carlo methods, to escape the inefficiency of exhaustive search.