Object`s e.g. human face, image base regression performing method, involves learning regression function for images, and applying learned function to predict entity that belongs to image without any declared entity

摘要

The method involves learning a regression function for images of an object by approximating the function as an additive function, where the object can be a human face or heart. Each image is declared an appropriate entity, which can be a patient age, position and anisotropic spreading of lung tumor or position of intracardial wall. The learned function is applied to predict the entity that belongs to the image without declared entity.