Head-pose invariant recognition of facial attributes

摘要

A system facilitates automatic recognition of facial expressions or other facial attributes. The system includes a data access module and an expression engine. The expression engine further includes a set of specialized expression engines, a pose detection module, and a combiner module. The data access module accesses a facial image of a head. The set of specialized expression engines generates a set of specialized expression metrics, where each specialized expression metric is an indication of a facial expression of the facial image assuming a specific orientation of the head. The pose detection module determines the orientation of the head from the facial image. Based on the determined orientation of the head and the assumed orientations of each of the specialized expression metrics, the combiner module combines the set of specialized expression metrics to determine a facial expression metric for the facial image that is substantially invariant to the head orientation.

主权项

1. A computer-implemented method for training a specialized recognition engine to recognize a preselected attribute of a facial image, the specialized recognition engine specialized for a range of head poses, the method comprising:
for each of a plurality of facial images of heads that are within the range of head poses for the specialized recognition engine:
identifying a pair comprising a known good facial image and a known good specialized recognition metric, wherein the known good facial image is for a same head as said facial image and has the same preselected attribute as said facial image but for a head pose within a range of head poses that is different than the range of head poses for the specialized recognition engine, and the known good specialized recognition metric is indicative of the preselected attribute of the known good facial image; andassociating a specialized recognition metric with said facial image, wherein the specialized recognition metric is derived from the known good specialized recognition metric; and training the specialized recognition engine by using pairs comprising the plurality of facial images and the associated specialized recognition metrics as a training set for supervised learning.