# US 7,609,891 B2 · Evaluation of element distribution within a collection of images based on pixel scatterness

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 7,609,891 B2
- Document ID: US7609891B2
- Status: Granted
- Kind: B2
- Application number: 11/217,942
- Filed: 2005-08-31
- Published: 2007-03-01
- Granted: 2009-10-27
- Earliest claimed: 2005-08-31
- Assignees: Sony Corporation; Sony Electronics Inc.
- Inventors: Pingshan Li
- Web page: https://patentgrep.com/patent/US7609891B2

## CPC

- G06V 10/42: Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V 30/2504: {Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches}
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method of evaluating element distribution within a data set, such as pixel characteristics whose scatterness is quantified. Scatterness is evaluated for individual scale levels as the image is merged through smaller scale levels. At each level, every pixel block (i.e., 2×2) in the region of interest merges to one pixel to form the next level. Select characteristics of the pixels are assessed in the merge, and the image may be simplified to fewer pixel states so the merges require less processor overhead. In one merge mode if any pixel in the block has the desired characteristic, then the resulting merged pixel is set with that characteristic. The merging rate of the pixels is used to determine the scatterness for the pixel distribution. A scatterness sequence can be used separately or in combination with known techniques for comparing images in applications such as search and classification.

## Claims

### Claim 1

A method of quantifying scatterness within images or video frames based on pixel characteristics, comprising:
(a) simplifying images in response to detection of desired pixel characteristics by setting each pixel of an image to a first state indicative of pixels lacking the desired characteristic or range of characteristics, or a second state indicative of pixels possessing the desired pixel characteristic or range of characteristics, so as to reduce possible states of each pixel;
(b) merging each block of pixels from within a plurality of pixel blocks within an image into a collection of single pixels having either said first state or said second state; said merging is performed in response to execution of programming within a computer processing device;
(c) registering the number or proportion of pixels having said second state within the pixels;
wherein every level in the merging and registration process is based on one of two pixel states;
(d) repeating steps (b) and (c) until a desired level of merging is reached; and
(e) generating scatterness sequence information in response to determining a merging rate of pixels of said second state for at least two of the merges;
wherein said quantifying of scatterness is performed within a classification or comparison of images or video frames.

### Claim 2 (depends on 1)

A method as recited in claim 1, wherein said images or video comprises at least image or video of at least two dimensions.

### Claim 3 (depends on 1)

A method as recited in claim 1, wherein said pixel characteristics comprise color characteristics, intensity characteristics, or a combination of color and intensity characteristics.

### Claim 4 (depends on 1)

A method as recited in claim 1, further comprising:
comparing the scatterness sequence for classifying the image according to said desired pixel characteristics;
wherein the scatterness sequence is either directly compared or compared after first compacting it by converting it to a logarithmic value.

### Claim 5 (depends on 1)

A method as recited in claim 1, further comprising:
comparing the scatterness sequence for classifying the image according to said desired pixel characteristic;
wherein the scatterness sequence is either directly compared or compared after first compacting it by converting it to a logarithmic value; and
comparing a rate of change in the scatterness sequence for classifying said desired pixel characteristics.

### Claim 6 (depends on 1)

A method as recited in claim 1, wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixel to have s selected characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or range of selected pixel characteristics, exceeds a predetermined threshold.

### Claim 7 (depends on 1)

A method as recited in claim 1:
wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixels to have a selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or range of selected pixel characteristics, exceeds a predetermined threshold; and
wherein said predetermined threshold comprises a condition in which at least one pixel in the block of pixels possesses the selected pixel characteristics.

### Claim 8 (depends on 1)

A method as recited in claim 1:
wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixels to have a selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or range of selected pixel characteristics, exceeds a predetermined threshold;
wherein merging of blocks of pixels into a collection of single pixels further comprises adjusting a state of adjacent pixels in response to the positioning of pixels having selected pixel characteristics within the block of pixels; and
wherein the state of adjacent pixels is adjusted to a first state indicative of pixels lacking selected pixel characteristic, or a second state indicative of pixels possessing the selected pixel characteristics.

### Claim 9 (depends on 1)

A method as recited in claim 1:
wherein merging of blocks of pixels into a collection of single pixels is performed according to,
f
i
+
1
⁡
[
m
,
n
]
=
0
if
{
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
+
1
]
=
0
;
f
i
⁡
[
m
,
n
]
=
1
if
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
1
;
f
i
+
1
⁡
[
m
,
n
]
=
1
if
{
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
⁢
+
⁢
1
,
⁢
2
⁢
⁢
n
]
⁢
=
⁢
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
⁢
2
⁢
⁢
n
⁢
+
⁢
1
]
⁢
=
⁢
1
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
,
2
⁢
⁢
n
]
=
1
;
⁢
f
⁢
i
+
1
⁡
[
m
,
n
]
=
1
,
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
2
⁢
⁢
n
⁢
+
2
]
=
0
,
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
,
2
⁢
⁢
n
+
2
]
=
0
}
if
{
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
2
⁢
⁢
n
]
=
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
,
2
⁢
⁢
n
]
=
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
⁢
,
2
⁢
⁢
n
+
1
]
=
1
;
⁢
f
⁢
i
+
1
⁡
[
m
,
n
]
=
1
,
⁢
f
⁢
i
[
2
⁢
⁢
m
+
2
,
2
⁢
⁢
n
⁢
]
=
0
,
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
2
,
2
⁢
⁢
n
+
1
]
=
0
}
if
{
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
2
⁢
⁢
n
]
=
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
,
2
⁢
⁢
n
]
=
1
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
⁢
,
2
⁢
⁢
n
+
1
]
=
0
;
⁢
f
⁢
i
+
1
⁡
[
m
,
n
]
=
1
,
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
,
2
⁢
⁢
n
⁢
+
2
]
=
0
,
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
2
,
2
⁢
⁢
n
+
1
]
=
0
,
f
i
⁡
[
2
⁢
m
+
2
,
2
⁢
n
+
2
]
=
0
}
if
{
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
2
⁢
⁢
n
]
=
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
,
2
⁢
⁢
n
+
1
]
=
0
,
⁢
⁢
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
⁢
,
2
⁢
⁢
n
]
=
0
,
f
⁢
i
⁡
[
2
⁢
⁢
m
+
1
⁢
,
2
⁢
⁢
n
+
1
]
=
1.
wherein fi[m,n] denotes the image pixel at position [m,n] at the ith level, while fi+1[m,n] represents the pixel as a result of the merge.

### Claim 10 (depends on 1)

A method as recited in claim 1, further comprising:
simplifying the image based on pixel characteristics by converting the image into an image having pixels in either a first state indicative of pixels lacking the desired characteristic, or a second state indicative of pixels possessing the desired pixel characteristics;
wherein the simplifying step is performed prior to merging of the pixels toward speeding up the merge process.

### Claim 11

A method of quantifying scatterness within images or video frames based on characteristics, comprising:
(a) merging each block of pixels from within a plurality of pixel blocks within an image into a collection of single pixels, which are set to either a first state indicative of pixels lacking a desired characteristic or range of characteristics, or a second state indicative of pixels possessing the desired pixel characteristic or range of characteristics;
said merging is performed in response to execution of programming within a computer processing device;
(b) registering the number or proportion of pixels having a desired pixel characteristic, or range of characteristics, within the pixels as a measure of scatterness for that characteristic;
wherein said pixel characteristics comprise color characteristics, intensity characteristics, or a combination of color and intensity characteristics;
(c) repeating steps (a) and (b) until a desired level of merging is reached; and
(d) generating scatterness sequence information in response to the registration of pixel characteristics for at least two of the merges;
wherein said quantifying of scatterness is performed within a classification or comparison of images or video frames.

### Claim 12 (depends on 11)

A method as recited in claim 11, further comprising; (e) comparing the scatterness sequence of the image with other images, or with classification thresholds for classifying the image according to the selected pixel characteristic.

### Claim 13 (depends on 11)

A method as recited in claim 11, further comprising
comparing the scatterness sequence of the image with other images, or with classification thresholds for classifying the image according to the selected pixel characteristic;
wherein said comparing comprises comparing a rate of change in the scatterness sequence for classifying the selected pixel characteristic.

### Claim 14 (depends on 11)

A method as recited in claim 11, wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixel to have the selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or a range of selected pixel characteristics, exceeds a predetermined threshold.

### Claim 15 (depends on 11)

A method as recited in claim 11:
wherein merging of blocks of pixels into a collection of single pixels comprises setting the mered pixel to have the selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or a range of selected pixel characteristics, exceeds a predetermined threshold; and
wherein said predetermined threshold comprises a condition in which at least one pixel in the block of pixels possesses the selected pixel characteristics.

### Claim 16 (depends on 11)

A method as recited in claim 11:
wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixel to have the selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or a range of selected pixel characteristics, exceeds a predetermined threshold;
wherein merging of blocks of pixels into a collection of single pixels further comprises adjusting a state of adjacent pixels in response to positioning of pixels having the selected characteristics within the pixel block; and
wherein the state is adjusted to a first state indicative of pixels lacking the selected pixel characteristic, or a second state indicative of pixels possessing the selected pixel characteristics.

### Claim 17

A method of classifying images or video frames based on quantifying scatterness sequences of pixel characteristics, comprising:
(a) merging each block of pixels from within a plurality of pixel blocks within an image into a collection of single pixels based on setting the resultant pixel to have either a first state indicative of pixels lacking the desired characteristic or range of characteristics, or a second state indicative of pixels possessing a selected pixel characteristic, or range of selected pixel characteristics, if the number of pixels within the pixel block that possess the selected pixel characteristic, or range of selected pixel characteristics, exceeds a predetermined threshold;
said merging is performed in response to execution of programming within a computer processing device;
(b) registering the number or proportion of pixels having said second state within the image or a selected portion of the image;
wherein said selected pixel characteristics comprise color characteristics, intensity characteristics, or a combination of color and intensity characteristics;
(c) merging blocks of pixels formed from said single pixels based on setting the resultant pixel to have either a first state indicative of pixels having said first state, or a second state indicative of pixels having said second state;
(d) registering the number or proportion of pixels having said second state, within the image or a selected portion of the image;
(e) repeating steps (c) and (d) until a selected number of merges are performed, or the resultant image size reaches a threshold value;
(f) generating scatterness sequence information in response to determining a merging rate of pixels of said second state;
wherein every level in the registration process is based on one of two pixel states; and
(g) comparing the scatterness sequence of the image with other images, or with classification thresholds for classifying the image according to that selected pixel characteristic.

### Claim 18 (depends on 17)

A method as recited in claim 17, further comprising comparing the scatterness sequence of the image with scatterness sequences of other images, or with scatterness classification thresholds for classifying the image according to that selected pixel characteristic.

### Claim 19 (depends on 17)

A method as recited in claim 17, wherein said
comparing comprises comparing a rate of change in the scatterness sequence for classifying the image characteristics.

### Claim 20 (depends on 17)

A method as recited in claim 17, wherein merging of blocks of pixels into a collection of single pixels comprises setting the merged pixel to have the selected pixel characteristic in response to detecting that the number of pixels possessing the selected pixel characteristic, or range of selected pixel characteristics, exceeds a predetermined threshold.

### Claim 21

A method of quantifying scatterness within images or video frames based on pixel characteristics, comprising:
(a) merging each block of pixels from within a plurality of pixel blocks within an image into a collection of single pixels;
said merging is performed in response to execution of programming within a computer processing device;
(b) registering the number or proportion of pixels having a desired pixel characteristic, or range of characteristics, within the pixels as a measure of scatterness for that characteristic;
(c) repeating steps (a) and (b) until a desired level of merging is reached; and
(d) generating scatterness sequence information in response to the registration of characteristics for at least two of the merges;
wherein said merging of blocks of pixels into a collection of single pixels is performed according to,
f
i
+
1
⁡
[
m
,
n
]
=
0
⁢
⁢
if
⁢
⁢
{
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
+
1
]
=
0
,
⁢
⁢
f
i
+
1
⁡
[
m
,
n
]
=
1
⁢
⁢
if
⁢
⁢
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
1
;
⁢
f
i
+
1
⁡
[
m
,
n
]
=
1
⁢
⁢
if
⁢
⁢
{
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
1
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
1
;
⁢
⁢
f
i
+
1
⁡
[
m
,
n
]
=
1
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
2
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
+
2
]
=
0
}
⁢
⁢
if
⁢
⁢
{
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
1
;
⁢
⁢
f
i
+
1
⁡
[
m
,
n
]
=
1
,
f
i
⁡
[
2
⁢
m
+
2
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
2
,
2
⁢
n
+
1
]
=
0
}
⁢
⁢
if
⁢
⁢
{
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
1
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
0
;
⁢
⁢
f
i
+
1
⁡
[
m
,
n
]
=
1
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
+
2
]
=
0
,
f
i
⁡
[
2
⁢
m
+
2
,
2
⁢
n
+
1
]
=
0
,
f
i
⁡
[
2
⁢
m
+
2
,
2
⁢
n
+
2
]
=
0
,
}
⁢
⁢
if
⁢
⁢
{
f
i
⁡
[
2
⁢
m
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
,
2
⁢
n
+
1
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
]
=
0
,
f
i
⁡
[
2
⁢
m
+
1
,
2
⁢
n
+
1
]
=
1
;
and
wherein fi[m, n] denotes the image pixel at position [m, n] at the ith level, while fi+1[m, n] represents the pixel as a result of the merge;
wherein said quantifying of scatterness is performed within a classification or comparison of images or video frames.