Infra-red contrast enhancement filter
A method and a system for enhancing contrast for an electro-optical video for live streaming is disclosed. An image block from an electro-optical frame of a video is received, where a number of pixels in the image block are identified to compute intensity of the image block. A distribution of intensity of the number of pixels in the image block is determined, and from the distribution, a degree of variation is determined, where a lower threshold and upper threshold for the intensity are determined. An upper set of pixels and a lower set of pixels is determined, which is then used to modify the distribution. The degree of variation is modified to determine uniformity in distribution intensity. A transform using the uniformity is constructed to enhance the contrast of the image block. The contrast of all image blocks of the electro-optical is enhanced by applying the transform.
1 . A method for enhancing contrast for an electro-optical video for live streaming, the method comprises:
receiving an image block, wherein:
the image block is part of an electro-optical frame,
the electro-optical frame comprises a plurality of image blocks,
the image block is one of the plurality of image blocks, and
the electro-optical frame is one of a plurality of electro-optical frames generated from an image sensor to form the electro-optical video;
computing intensity for the image block, wherein the intensity for the image block is computed based on the number of pixels contained therein;
determining a distribution of intensity of the number of pixels in the image block, wherein each of the number of pixels has a pixel intensity;
quantizing a degree of variation in the intensity for the number of pixels, wherein quantizing the degree of variation in the intensity for the number of pixels includes calculatin a/the Gini coefficient by
G
=
∑
x
(
U
(
x
)
-
C
(
x
)
)
∑
x
U
(
x
)
,
where G is Gini coefficient, U(x) is a uniform distribution function and C(x) is a cumulative distribution function;
determining an upper threshold and a lower threshold for intensity of the number of pixels based on the degree of variation, wherein the lower threshold is determined by:
minThd
=
totalPixelNumber
numberOfHistgramBins
,
and the upper threshold is determined by maxThd=(1−giniCoef)*maxHistogramBinCount+giniCoef*minThd, and wherein the upper threshold is a weighted sum of maximum number of pixels in a histogram bin and the lower threshold;
determining an upper set of the number of pixels outlying the upper threshold;
determining a lower set of the number of pixels outlying the lower threshold;
modifying the distribution of pixel intensity by:
for the upper set, reducing the pixel intensity of each of the pixels to a first predetermined intensity, and
for the lower set, increasing the pixel intensity to a second predetermined intensity;
modifying the degree of variation to determine a uniformity in distribution of intensity for the image block;
constructing a transform using the uniformity to enhance the contrast of the image block; and
applying the transform to the plurality of image blocks to enhance the contrast of the plurality of image blocks.
2 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein the image block has a size of 320*180 pixels.
3 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein each of the plurality of image blocks are separate blocks sharing no common area of the electro-optical frame.
4 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein the plurality of electro-optical frames is generated through illumination.
5 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein the plurality of electro-optical frames is generated through thermal imaging.
6 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein determining the distribution of intensity of the number of pixels in the image block includes generating a histogram for the number of pixels, and wherein the histogram is generated by hist[pixVal]=hist[pixVal]+1.
7 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 6 , wherein the method further comprises:
computing function of distribution for the histogram;
normalizing the histogram; and
upon normalizing the histogram, determining weight of the histogram by calculating an inverse Euclidean distance between pixels and center of the histogram,
wherein the weight is determined by
w
k
i
=
1
(
x
i
-
x
k
)
2
+
(
y
i
-
y
k
)
2
+
ε
.
8 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 7 , wherein the method further comprises:
computing energy for each pixel using the function of distribution:
x
k
′
=
CDF
k
(
x
)
;
and
assigning weight of the histogram based on the energy by
x
″
=
∑
k
x
k
′
·
w
k
.
9 . The method for enhancing contrast for the electro-optical video for live streaming as claimed in claim 1 , wherein the first predetermined intensity corelates to the upper threshold and the second predetermined intensity corelates to the lower threshold.
10 . A system for enhancing contrast for an electro-optical video for live streaming, the system comprising: an image processor to: receive an image block, wherein: the image block is part of an electro-optical frame, the electro-optical frame comprises a plurality of image blocks, the image block is one of the plurality of image blocks, and the electro-optical frame is one of a plurality of electro-optical frames generated from an image sensor to form the electro-optical video; compute intensity for the image block, wherein for the image block, the intensity is computed based on a number of pixels therein; and determine a distribution of intensity of the number of pixels in the image block, wherein each of the number of pixels has a pixel intensity; a quantizer communicably coupled to the image processor to: quantize a degree of variation in the intensity for the number of pixels, wherein the degree of variation in the intensity for the number of pixels by calculating a Gini coefficient by
G
=
∑
x
(
U
(
x
)
-
C
(
x
)
)
∑
x
U
(
x
)
,
where G is Gini coefficient, U(x) is a uniform distribution function and C(x) is a cumulative distribution function; determine an upper threshold and a lower threshold for intensity of the number of pixels based on the degree of variation, wherein the lower threshold is determined by:
minThd
=
totalPixelNumber
numberOfHistgramBins
,
and the upper threshold is determined by m axThd=(1−giniCoef)*maxHistogramBinCount+giniCoef*minThd, and wherein the upper threshold is a weighted sum of maximum number of pixels in a histogram bin and the lower threshold; determining an upper set of the number of pixels outlying the upper threshold; determining a lower set of the number of pixels outlying the lower threshold; modifying the distribution of pixel intensity by: for the upper set, reducing the pixel intensity of each of the pixels to a first predetermined intensity, and for the lower set, increasing the pixel intensity to a second predetermined intensity; modifying the degree of variation to determine a uniformity in distribution of intensity for the image block; constructing a transform using the uniformity to enhance the contrast of the image block; and applying the transform to the plurality of image blocks to enhance the contrast of the plurality of image blocks.
11 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein the image block has a size of 320*180 pixels.
12 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein each of the plurality of image blocks are/is separate blocks sharing no common area of the electro-optical frame.
13 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein the plurality of electro-optical frames is generated through illumination.
14 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein the plurality of electro-optical frames is generated through thermal imaging.
15 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein the image processor is configured to determine the distribution of intensity of the number of pixels in the image block by generating a histogram for the number of pixels, and wherein the histogram is generated by hist[pixVal]=hist[pixVal]+1.
16 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 15 , wherein the image processor is configured to:
compute function of distribution for the histogram;
normalize the histogram; and
upon normalizing the histogram, determine the weight of the histogram by calculating an inverse Euclidean distance between pixels and the center of the histogram,
wherein the weight is determined by
w
k
i
=
1
(
x
i
-
x
k
)
2
+
(
y
i
-
y
k
)
2
+
ε
.
17 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 16 , wherein the image processor is configured to:
compute energy for each pixel using the function of distribution:
x
k
′
=
CDF
k
(
x
)
;
and
assign a/the weight of the histogram based on the energy by
x
″
=
∑
k
x
k
′
·
w
k
.
18 . The system for enhancing contrast for the electro-optical video for live streaming as claimed in claim 10 , wherein the first predetermined intensity correlates to the upper threshold and the second predetermined intensity correlates to the lower threshold.