Multi-cue object detection and analysis

Patent metadata

Field	Value
Publication number	US-9396548-B2
Application number	US-201514861085-A
Country	US
Kind code	B2
Filing date	Sep 22, 2015
Priority date	Jun 14, 2012
Publication date	Jul 19, 2016
Grant date	Jul 19, 2016

Abstract

Foreground objects of interest are distinguished from a background model by dividing a region of interest of a video data image into a grid array of individual cells. Each of the cells are labeled as foreground if accumulated edge energy within the cell meets an edge energy threshold, or if color intensities for different colors within each cell differ by a color intensity differential threshold, or as a function of combinations of said determinations.

First claim

What is claimed is:

1. A computer-implemented method for distinguishing foreground objects of interest from a background model, the method comprising executing on a processing unit the steps of:
dividing a region of interest of a video data image into a grid array of a plurality of individual cells;
acquiring frame image data for each of the cells;
determining a first background indication for each of the cells that have determined color intensities that do not exceed others of the determined color intensities for the cell by a color intensity differential threshold;
determining a first foreground indication for each of the cells that have one of the determined color intensities greater than another of the determined color intensities for that cell by the color intensity differential threshold;
determining a second background indication for each of the cells that have an accumulated energy of edges detected within the cells that less than an edge energy threshold;
determining a second foreground indication for each of the cells that have an accumulated energy of edges detected within the cells that meets or exceeds the edge energy threshold;
labelling as foreground or background each of the cells in response to applying a combination rule to the foreground indications and the background indications for the cells; and
using the frame image data from the cells labeled as foreground to define a foreground object.

2. The method of claim 1 , further comprising:
integrating computer-readable program code into a computer system comprising the processing unit, a computer readable memory and a computer readable hardware storage device; and
wherein the computer readable program code is embodied on the computer readable hardware storage device and comprises instructions for execution by the processing unit via the computer readable memory that cause the processing unit to perform the steps of dividing the region of interest into the grid array, acquiring the frame image data for each of the cells, determining the first background indication for each of the cells, determining the first foreground indication for each of the cells, determining the second background indication for each of the cells, determining the second foreground indication for each of the cells, labelling as foreground or background each of the cells in response to applying the combination rule, and using the frame image data from the cells labeled as foreground to define the foreground object.

3. The method of claim 1 , further comprising:
determining a third background indication for each of the cells that have sums of the determined color intensities that exceed sums of the determined color intensities of adjacent neighborhood cells by a color consistency threshold value; and
determining a third foreground indication for each of the cells that have sums of the determined color intensities that do not exceed sums of the determined color intensities of adjacent neighborhood cells by the color consistency threshold value.

4. The method of claim 3 , further comprising:
differentially weighting the first foreground indication, the second foreground indication and the third foreground indication relative to each other; and
wherein the step of labelling as foreground or background each of the cells in response to applying the combination rule to the foreground indications and the background indications for the cells comprises labelling the cells that have less than all of the first foreground indication, the second foreground indication and the third foreground indication as foreground in response to having a one of the first foreground indication, the second foreground indication and the third foreground indication that is more heavily weighted than others of the first foreground indication, the second foreground indication and the third background indication.

5. The method of claim 3 , wherein the step of labelling as foreground or background each of the cells in response to applying the combination rule to the foreground indications and the background indications for the cells comprises:
labelling the cells that have at least two of the first foreground indication, the second foreground indication and the third foreground indication as foreground; and
labelling the cells that do not have at least two of the first foreground indication, the second foreground indication and the third foreground indication as background.

6. The method of claim 5 , wherein the step of labelling as foreground or background each of the cells in response to applying the combination rule to the foreground indications and the background indications for the cells comprises:
labelling the cells that have all of the first foreground indication, the second foreground indication and the third foreground indication as foreground; and
labelling the cells that do not have all of the first foreground indication, the second foreground indication and the third foreground indication as background.

7. The method of claim 3 , further comprising:
applying a Sobel operator to each cell to create an edge patch; and
determining for each cell the second foreground indication in response to determining a value of “1”, and the second background indication in response to determining a value of “0” according to:

f
e

⁡

(

c
m

)

=

{

1

∑

i
=
1

N

⁢

⁢

E
⁡

(

x
i

)

>

t
e

0

∑

i
=
1

N

⁢

⁢

E
⁡

Assignees

• IBM

Inventors

• Datta Ankur
• Feris Rogerio S
• Pankanti Sharathchandra U
• Wang Xiaoyu

Classifications

• G06F2218/12

  Classification; Matching · CPC title

• G06V10/50

  by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis · CPC title

• G06V10/28

  Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns · CPC title

• G06V10/255

  Detecting or recognising potential candidate objects based on visual cues, e.g. shapes · CPC title

• G06K9/3241

  Physics · mapped topic