Method for locating a subject's lips in a facial image

• US 5,805,745 A
• Filed: 06/26/1995
• Issued: 09/08/1998
• Est. Priority Date: 06/26/1995
• Status: Expired due to Term

• Current Assignee: Alcatel-Lucent USA, Inc. (Nokia Corporation)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Graf, Hans Peter
• Primary Examiner(s): SHALWALA, BIPIN H
• Application Number:
  US08/494,900
• Time in Patent Office: 1,170 Days
• Field of Search: 382/115, 382/117, 382/118, 382/190, 382/191, 382/192, 382/194, 382/195, 382/279, 382/286, 382/308, 382/155, 382/156, 382/159, 382/260-264, 382/205, 382/282, 382/224, 382/170, 382/291, 348/14, 348/15, 348/17, 348/18, 348/19, 348/20, 348/22, 348/389
• US Class Current: 382/291
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation
• CPC Class Codes: H04N 19/20 using video object coding
• CPC Class Codes: H04N 7/142 Constructional details of the terminal equipment, e.g. arrangements of the camera and the display

Abstract

An embodiment of the present invention locates facial features in an image by bandpass filtering the image and then performing morphological operations followed by a thresholding operation. This initial processing identifies candidate areas where facial features may be located. The candidate areas are evaluated by classifiers to determine if a facial feature, such as an eye or mouth, has been located.

Claims

1. A method for locating a subject'"'"'s lips in a image, comprising the steps of:

bandpass filtering the image to produce a bandpass filtered image;

morphologically processing the bandpass filtered image to produce an enhanced image by convolving the bandpass filtered image with a rectangular shape;

thresholding the enhanced image to form a binary image having a plurality of connected components; and

using a classifier to identify at least one connected component corresponding to the subject'"'"'s mouth;

determining an intensity profile along a vertical strip across the image of the subject'"'"'s mouth;

determining an intensity variation of the intensity profile; and

comparing the intensity profile to a plurality of stored intensity profiles to locate an edge of at least one of the lips if the intensity variation is below a first threshold, and comparing the intensity profile to a second threshold to locate the edge of at least one of the lips if the intensity variation is above the first threshold.

2. The method of claim 1, wherein the step of morphologically processing comprises convolving the bandpass filtered image with a rectangular shape having vertical and horizontal dimensions similar to the subject'"'"'s mouth.

3. The method of claim 1, wherein the step of bandpass filtering comprises lowpass filtering the image to produce a lowpass filtered image, highpass filtering the lowpass filtered image to produce a highpass filtered image, and subtracting the highpass filtered image from the lowpass filtered image to produce the bandpass filtered image.

4. The method of claim 1, wherein the step of using a classifier comprises using a linear classifier.

5. The method of claim 1, wherein the step of using a classifier comprises using a neural network.

6. The method of claim 1, wherein the step of bandpass filtering comprises convolving the image with a rectangular shape.