Automatic red-eye object classification in digital images using a boosting-based framework
Automatic red-eye object classification in digital images using a boosting-based framework. In a first example embodiment, a method for classifying a candidate red-eye object in a digital photographic image includes several acts. First, a candidate red-eye object in a digital photographic image is selected. Next, a search scale set and a search region for the candidate red-eye object where an eye object may reside is determined. Then, the number of subwindows that satisfy an AdaBoost classifier is determined. This number is denoted as a vote. Next, the maximum size of the subwindows that satisfy the AdaBoost classifier is determined. Then, a normalized threshold is calculated by multiplying a predetermined constant threshold by the calculated maximum size. Next, the vote is compared with the normalized threshold. Finally, the candidate red-eye object is transformed into a true red-eye object if the vote is greater than the normalized threshold.
1. A method for classifying a candidate red-eye object in a digital image, the method comprising the acts of:
i) selecting a candidate red-eye object in a digital image;
ii) determining the number of subwindows of the selected candidate red-eye object passed through an AdaBoost classifier for a certain classification stage;
iii) calculating a ratio between either:
iii.a) the number of subwindows passed through a current stage and the number of subwindows passed through a previous stage; or
iii.b) the number of subwindows passed through the current stage and the difference between the total number of subwindows to be tested by the AdaBoost classifier and the number of subwindows passed through the previous stage;
iv) determining that the ratio is greater than or equal to a stage threshold; and
v) transforming the candidate red-eye object into a true red-eye object.
2. The method as recited in claim 1 , wherein the acts iii.a), iv) and v) are performed according to the following equation:
S
i
=
{
1
if
V
i
α
/
V
i
(
α
-
1
)
≥
η
α
;
α
=
1
,
2
,
…
,
λ
′
0
otherwise
where:
S i =1 denotes that the candidate red-eye object is a true red-eye object;
S i =0 denotes that the candidate red-eye object is a false red-eye object;
η α is a predetermined parameter;
λ′ denotes the number of stages for which the classification is performed; and
V i + is the number of subwindows passed the αth stage of AdaBoost.
3. The method as recited in claim 1 , wherein the acts iii.b), iv) and v) are performed according to the following equation:
S
i
=
{
1
if
V
i
1
/
V
i
T
≥
η
1
1
if
V
i
α
/
(
V
i
T
-
V
i
(
α
-
1
)
)
≥
η
α
;
α
=
2
,
3
,
…
,
λ
′
0
otherwise
where:
S i =1 denotes that the candidate red-eye object is a true red-eye object;
S i =0 denotes that the candidate red-eye object is a false red-eye object;
η α , for α=1, 2, 3, . . . , λ′, is a predetermined parameter;
λ′ denotes the number of stages for which the classification is performed;
V i T is the total number of subwindows to classify; and
V i α is the number of subwindows that passed the αth stage of AdaBoost.
4. The method as recited in claim 1 , further comprising the act of:
viii) transforming the true red-eye object into a corrected eye object by removing the red-eye effect from the true red-eye object.
5. One or more non-transitory computer-readable media having computer-readable instructions thereon which, when executed by a processor, implement a method for classifying a candidate red-eye object in a digital image, the method comprising the acts of:
i) selecting a candidate red-eye object in a digital image;
ii) determining the number of subwindows of the selected candidate red-eye object passed through an AdaBoost classifier for a certain classification stage;
iii) calculating a ratio between either:
iii.a) the number of subwindows passed through a current stage and the number of subwindows passed through a previous stage; or
iii.b) the number of subwindows passed through the current stage and the difference between the total number of subwindows to be tested by the AdaBoost classifier and the subwindows passed through the previous stage;
iv) determining that the ratio is greater than or equal to a stage threshold; and
v) transforming the candidate red-eye object into a true red-eye object.
6. An image processing apparatus comprising:
an electronic display; and
a processor in electronic communication with the electronic display; and
the one or more computer-readable media as recited in claim 5 .
7. The image processing apparatus as recited in claim 6 , wherein:
the image processing apparatus comprises a printer;
the one or more computer-readable media comprises one or more of a RAM, a ROM, and a flash EEPROM; and
the electronic visual display comprises a liquid crystal display.