Methods for discriminating moving objects in motion image sequences

• US 7,639,874 B2
• Filed: 08/03/2006
• Issued: 12/29/2009
• Est. Priority Date: 08/03/2006
• Status: Active Grant

• Current Assignee: Innovation Asset Collective
• Original Assignee: Tandent Vision Science, Inc.
• Inventors: Bushell, Steven Joseph, Friedhoff, Richard Mark, Maxwell, Bruce Allen
• Primary Examiner(s): Seth; Manav
• Application Number:
  US11/498,621
• Publication Number:
  US 20080031526A1
• Time in Patent Office: 1,244 Days
• Field of Search: 382/103, 382128-131, 382/162, 382/164, 382/171, 382/173, 382/180, 382/181, 382224-225, 348169-172
• US Class Current: 382/173
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation
• CPC Class Codes: G06T 7/215 Motion-based segmentation
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
• CPC Class Codes: G06V 10/422 for representing the structure of the pattern or shape of an object therefor
• CPC Class Codes: G06V 10/457 by analysing connectivity, e.g. edge linking, connected component analysis or slices
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 10/60 relating to illumination properties, e.g. using a reflectance or lighting model

Abstract

In an exemplary embodiment of the present invention, an automated, computerized method is provided for classifying pixel values in a motion sequence of images. According to a feature of the present invention, the method comprises the steps of determining spectral information relevant to the sequence of images, and utilizing the spectral information to classify a pixel as one of background, shadow and object.

Claims

1. An automated, computerized method for classifying pixel values in a motion sequence of images, comprising the steps of:

a computer executing the following steps;

determining spectral information relevant to the sequence of images, the spectral information including a characteristic spectral ratio for the background; and

utilizing the spectral information to classify a pixel as one of background, shadow and object.

2. The method of claim 1, wherein the step of utilizing the spectral information to classify a pixel as one of background, shadow and object is carried out by:

determining a characteristic spectral ratio for a static background of the sequence of images;

calculating, for a subject pixel of an image in the sequence, a spectral ratio between the subject pixel and a corresponding pixel of the static background; and

comparing the spectral ratio for the subject pixel to the characteristic spectral ratio.

3. The method of claim 2, wherein the spectral ratio for the subject pixel is a normalized spectral ratio.

4. A computer system comprising:

a CPU; and

a memory storing a sequence of images;

the CPU arranged and configured to execute a routine to determine spectral information relevant to the sequence of images, wherein the spectral information comprises a characteristic spectral ratio for the background, and utilize the spectral information to classify a pixel as one of background, shadow and object.

5. An automated, computerized method for classifying pixel values in a motion sequence of images, comprising the steps of:

a computer executing the following steps;

determining spectral information relevant to the sequence of images, the spectral information including a BIDR cylinder representation of pixel values, in a color space; and

utilizing the spectral information to classify a pixel as one of background, shadow and object.

6. The method of claim 5, wherein the step of utilizing the spectral information to classify a pixel as one of background, shadow and object is carried out by:

generating a BIDR cylinder representative of values of a subject pixel location throughout the sequence of images; and

comparing a value of the subject pixel location in one image, to the BIDR cylinder.

7. The method of claim 5, wherein the color space is an RGB space.

8. The method of claim 6, wherein the BIDR cylinder representative of values of a subject pixel location throughout the sequence of images, comprises a temporal linear token of pixel values at the location.

9. The method of claim 6, wherein the step of generating a BIDR cylinder representative of values a subject pixel location throughout the sequence of images is carried out by calculating state probabilities for pixel values at the subject pixel location, and generating the BIDR cylinder representation as a function of the state probabilities.

10. The method of claim 9, wherein the step of calculating state probabilities for pixel values at the subject pixel location is carried out by utilizing a Gaussian mixture model.

11. A computer system comprising:

a CPU; and

a memory storing a sequence of images;

the CPU arranged and configured to execute a routine to determine spectral information relevant to the sequence of images, wherein the spectral information comprises a BIDR cylinder representation of pixel values, in a color space, and utilize the spectral information to classify a pixel as one of background, shadow and object.

12. A computer program product, disposed on a computer readable media, the product including computer executable process steps operable to control a computer to:

provide a sequence of image files, each depicting an image, in a computer memory;

determine spectral information relevant to the sequence of images, wherein the spectral information comprises a characteristic spectral ratio for the background; and

utilize the spectral information to classify a pixel as one of background, shadow and object.

13. A computer program product, disposed on a computer readable media, the product including computer executable process steps operable to control a computer to:

provide a sequence of image files, each depicting an image, in a computer memory;

determine spectral information relevant to the sequence of images, wherein the spectral information comprises a BIDR cylinder representation of pixel values, in a color space; and

utilize the spectral information to classify a pixel as one of background, shadow and object.