US 7,627,149 B2Granted

Dot templates for object detection in images

AssigneeQuantum Signal, LLC

InventorsVictor E. Perlin, Mitchell M. Rohde, Robert M. Lupa

Life of this patent

Estimating the term…

Loading patent…

Abstract

Dot templates are used for detecting objects within images. A computer-implemented method is performed for each of a number of dot templates corresponding to the object to be detected within an image. Each dot template is defined as a collection of points. At each position within the image, a value of the image is determined at each point of the dot template. The dot template is effectively overlaid at the given position within the image. A score of the dot template at this position is determined, based on the values of the image determined at the points of the dot template. Where the score is greater than a predetermined threshold, it can be concluded that the object is at least potentially located within the image at the position in question at which the dot template has been effectively overlaid.

Claims12

Dependent claims

A computer-readable medium having software stored thereon, execution of the software by hardware causing a method to be performed, the method comprising:

for each dot template of one or more dot templates corresponding to an object to be detected within an image, where each dot template is defined as a collection of points,
at each position of one or more positions within the image,
determining a value of the image at each point of the dot template at the position of the image, such that the dot template is effectively overlaid at the position;
determining a score of the dot template at the position based on the values of the image determined at the points of the dot template;
and,
where the score of the dot template is greater than a predetermined threshold, concluding that the object is at least potentially located within the image at the position at which the dot template has been effectively overlaid,
wherein the object to which the dot templates corresponds includes a face feature and the image is a face,
and wherein the face feature is a face edge, and determining the score of the dot template at the position based on the values of the image determined at the points of the dot template comprises determining the score as
∑
pnt
⁢
⁢
pairs
⁢
abs
⁡
(
I
1
-
I
2
)
(
I
1
+
I
2
)
,
for a number of image value point pairs I1, I2, where I1 of each image value point pair is a value of the image at a point corresponding to a face-side of a potential face edge of the face and I2 of each image value point pair is a value of the image at a point corresponding to a non-face-side of the potential face edge of the face.

29 more elements

A computer-readable medium having software stored thereon, execution of the software by hardware causing a method to be performed, the method comprising:

for each dot template of one or more dot templates corresponding to an object to be detected within an image, where each dot template is defined as a collection of points,
at each position of one or more positions within the image,
determining a value of the image at each point of the dot template at the position of the image, such that the dot template is effectively overlaid at the position;
determining a score of the dot template at the position based on the values of the image determined at the points of the dot template;
and,
where the score of the dot template is greater than a predetermined threshold, concluding that the object is at least potentially located within the image at the position at which the dot template has been effectively overlaid,
wherein the object to which the dot templates corresponds includes a face feature and the image is a face, and
wherein the face feature is selected from the group comprising eyes, eyebrows, and nostrils, and determining the score of the dot template at the position based on the values of the image determined at the points of the dot template comprises determining the score as
∑
pnt
⁢
⁢
pairs
⁢
(
I
1
-
I
2
)
(
I
1
+
I
2
)
,
for a number of image value point pairs I1, I2, where I1 of each image value point pairs is a value of the image at a point corresponding to a face feature-side of a potential face feature and I2 of each image value point pair is a value of the image at a point corresponding to a non-face feature-side of the potential face feature.

27 more elements

A computer-readable medium having software stored thereon, execution of the software by hardware causing a method to be performed, the method comprising:

for each dot template of one or more dot templates corresponding to an object to be detected within an image, where each dot template is defined as a collection of points,
at each position of one or more positions within the image,
determining a value of the image at each point of the dot template at the position of the image, such that the dot template is effectively overlaid at the position;
determining a score of the dot template at the position based on the values of the image determined at the points of the dot template;
and,
where the score of the dot template is greater than a predetermined threshold, concluding that the object is at least potentially located within the image at the position at which the dot template has been effectively overlaid,
wherein the object to which the dot templates corresponds includes a face feature and the image is a face,
and wherein the face feature comprises lips, and determining the score of the dot template at the position based on the values of the image determined at the points of the dot template comprises determining the score as
∑
pnt
⁢
⁢
pairs
⁢
g
1
-
g
2
,
for a number of point pairs I1, I2, where I1 of each point pair is a point corresponding to a lips-side of potential lips and I2 of each point pair is a point corresponding to a non-lips-side of the potential lips,
g
X
=
G
X
R
X
+
G
X
+
B
X
,
and Rx, Gx, Bx are red, green, and blue values of the image at the point Ix.

33 more elements

Dependencies are inferred from the claim text.

Facts

Application: 11/286,457Filed 25 Nov 2005
Published: US 2007/0122008 A131 May 2007
Granted: 1 Dec 2009
Term ends: Not estimatedThe source records don’t support an estimate.

Classification7 codes

G06V 40/165
Detection; Localisation; Normalisation: using facial parts and geometric relationships
G06T 7/74
Using feature-based methods: involving reference images or patches
G06V 10/757
Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g.…:…
G06V 40/161
Human faces, e.g. facial parts, sketches or expressions: detection; Localisation; Normalisation