Search International and National Patent Collections

1. (WO2017151759) CATEGORY DISCOVERY AND IMAGE AUTO-ANNOTATION VIA LOOPED PSEUDO-TASK OPTIMIZATION

Pub. No.:    WO/2017/151759    International Application No.:    PCT/US2017/020185
Publication Date: Sat Sep 09 01:59:59 CEST 2017 International Filing Date: Thu Mar 02 00:59:59 CET 2017
IPC: G06N 3/04
G06N 3/08
Applicants: THE UNITED STATES OF AMERICA, AS REPRESENTED BY THE SECRETARY, DEPARTMENT OF HEALTH AND HUMAN SERVICES
Inventors: LU, Le
WANG, Xiaosong
SUMMERS, Ronald, M.
Title: CATEGORY DISCOVERY AND IMAGE AUTO-ANNOTATION VIA LOOPED PSEUDO-TASK OPTIMIZATION
Abstract:
Methods and apparatus are disclosed for providing a looped deep pseudo-task automation approach for automatic category discovery as can be applied to a collection of images. In one example of the disclosed technology, a method of analyzing a collection of images includes extracting at least a portion of the activation values and weights associated with internal nodes of the neural network responsive to a respective input image of the collection of images being applied to the neural network, encoding the extracted activation values and weights, producing encoded vectors, clustering at least a portion of the collection of images based on similarities of the encoded vectors, to produce a plurality of clusters, and evaluating the clusters with a convergence criteria.