This document summarizes various algorithms that can be used for human organ localization in medical images. It discusses methods such as using probabilistic patch-based label fusion with image registration to localize the left ventricle, global-to-local regression forests to localize multiple organs, and using marginal and full space learning with coarse and fine levels to localize the left ventricle. Additionally, it covers using probabilistic edge maps generated from structured decision forests to localize the left ventricle and atrium, as well as using random decision forests and 3D visual features to detect anatomical structures in CT volumes. The document analyzes approaches such as combining random forests with Hough regression and Markov random fields to localize complex anatomical