Seminar on Computational Learning and Adaptation



Rotation Invariant Neural Network-Based Face Detection


Shumeet Baluja
Justsystem Pittsburgh Research Center/CMU
Pittsburgh, PA
baluja@jprc.com



We will present a state-of-the-art neural network-based face detection system. Unlike similar systems, this system is not limited to detecting upright, frontal faces; our system detects faces at any degree of rotation in the image plane. The system employs multiple networks. First, ``router'' networks are used to determine each input window's orientation. This information is used to transform the image into a standard orientation for input into multiple ``detector'' networks. The training methods and analyses for both types of networks are described in detail. We will present empirical results on several large test sets and discuss ideas for extending this work to out-of-plane rotations.

Joint work with Henry Rowley (CMU) and Takeo Kanade (CMU).

Relevant Papers and an On-Line Demo can be found at: http://www.cs.cmu.edu/~baluja 

Date: Thurs., April 16; Time: 4:15-5:30PM; Place: Gates 100

The goal of this seminar is to increase communication among local researchers with interests in computational approaches to learning and adaptation. If you would like to be added to (or removed from) the mailing list, or if you are interested in giving a talk in the seminar, please send email to iba@isle.org.


Return to seminar schedule.