Rotation Invariant Neural-Network Based Face Detection

1
Rotation Invariant Neural-Network Based Face
Detection
2
Overview

• Multiple Neural Networks
• Router Networks
• Detector Networks

3

• Overview of how the algorithm works

4
Input and output of the router network
5
Rotation NetworkOutputs are generated as
weighted vectors

• Average of the weighted vectors is interpreted as
  an angle
• 1048 training images labeled by face, eyes, tip
  of the nose, corners and centers of the mouth
• Each training face is rotated 15 times in a
  circle

6
Rotation Neural Net Description

• 400 layers on the input layer (20X20)
• Hidden layer of 15 units, output layer of 36
  units.
• Hyperbolic tangent activation function
• Standard error back propigation

7
Detector Network

• Identical to the routing network.
• Trained by positive (contains faces) and negative
  images (does not contain faces).
• Weights are initially random for the first
  training iteration.
• Training on non-face images, add false positives
  to the non-image

8
Adding False Positives to the training set as
negative images
9
Arbitration Scheme

• Detection of Different Faces at different angles
  in the same image
• Detections are placed in 4 dimensional space -
  x,y,angle, pyramid level, quantized in 10 degree
  increments.
• Two independently trained networks are ANDed to
  improve the success rate.

10
(No Transcript)
11
Empirical Results

• 130 images, 511 faces

12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
Sample Images
16
(No Transcript)
17
Conclusions

• Represents ways of integration multiple neural
  nets
• Speed of implementation
• Face Detection VS Facial Recognition