EXTENDING DATA-DRIVEN DETECTION TO PREDICTION OF OBJECT PART LOCATIONS

摘要

PROBLEM TO BE SOLVED: To provide methods and systems for detecting an object part location on the basis of an extended data-driven detection.SOLUTION: A specific relevance value between configurations of parts with respect to a set of training images annotated with annotating objects is defined. A similarity learning algorithm is applied with respect to the parts to obtain a similarity function on the basis of the similarity between the part configurations. The similarity learning algorithm receives a set of a positive pair having similar part configuration and a negative pair having different configuration and returns the similarity function that tends to assign a high score to a new positive pair and a low score to a negative pair. A similarity between a new image and the training images can be measured by utilizing the learned similarity function to obtain a neighbor image, and a visible and/or non-visible part location with respect to the image can be predicted based on the neighbor image.