Image processing apparatus and method for performing nonlinear gamma corrections

Claim 2

An image processing apparatus for performing nonlinear gamma corrections to a plurality of image signals for a single picture frame, comprising: image signal sampling means for sampling the plurality of image signals of said single picture frame and producing a sampling means output corresponding to the sampled image signals, learning type gamma selector means, coupled to said image signal sampling means, for: a) establishing, in advance, a nonlinear relation between sampling means outputs under various conditions and a plurality of nonlinear gamma curves; b) judging said sampling means output according to said established nonlinear relation; and c) selecting one of the plurality of nonlinear gamma curves for correction; and nonlinear gamma correction means, coupled to said learning type gamma selector means, for correcting said plurality of image signals using the selected nonlinear gamma curve.

Claim 3

An image processing apparatus according to claim 1, wherein the learning type gamma selector means comprises: a plurality of first and second recognition units interlinked to form a network having a plurality of tree-branch structures, each of said plurality of tree-branch structures having a plurality of layers wherein one of said plurality of layers is designated as a front layer, wherein said plurality of second recognition units are situated on the front layer, each of said plurality of first recognition units comprising: a signal inputting unit for receiving said sampling means output, a quantizer, coupled to said signal inputting unit, for receiving and quantizing said image signals received by said signal inputting unit, a channel selector unit, coupled to said quantizer and having at least one channel input terminal and at least one channel output terminal, wherein channels are selected according to the output of said quantizer, wherein each of said plurality of second recognition units comprises at least one channel input terminal, at least one channel output terminal linked to said at least one channel input terminal, wherein said linkage has a varying strength, and a channel learning unit, for receiving a teaching signal and, according to said teaching signal, modifying the linkage strength between said at least one channel input terminal and the at least one channel output terminal.

Claim 4

An image processing apparatus according to claim 1, wherein said image signal sampling means comprises a picture splitting unit for splitting said plurality of image signals into a plurality of picture blocks and a characteristic signal sampling unit for sampling said plurality of picture blocks and producing an output corresponding thereto.

Claim 5

An image processing apparatus according to claim 3, wherein said characteristic signal sampling unit comprises a brightness sampling unit for sampling brightness signals from said picture blocks, each one of said plurality of picture blocks having a maximum, a medium and a minimum brightness value; and a brightness computation unit for computing the maximum, the medium and the minimum brightness values within each of said picture blocks.

Claim 6

An image processing apparatus according to claim 4, wherein said brightness sampling unit comprises a brightness signal converting means for composing brightness signals from the image signals.

Claim 7

An image processing apparatus according to claim 3, wherein said established nonlinear relation is between said characteristic signal sampling unit output and said plurality of nonlinear gamma curves.

Claim 8

An image processing apparatus according to claim 1, wherein said image signal sampling means comprises: histogram computation processing means for a) inputting and processing the image signals of one picture frame, and b) obtaining and outputting a respective frequency for each of a plurality of signal levels for the processed image signals.

Claim 9

An image processing apparatus according to claim 7, wherein said histogram computation processing means comprises plural signal level comparators, each of said comparators for comparing the levels of picture information with a level having a certain spread in its value and outputting a comparison result with a digit number of 1 or 0, and plural counters for counting the output of said signal level comparators.

Claim 10

An image processing apparatus according to claim 7, wherein said histogram computation processing means comprises a brightness sampling unit for sampling brightness signals, plural signal level comparators, each of said comparators for comparing the brightness signals sampled through said brightness sampling unit with a level having a certain spread in its value and outputting a comparison result with a digit number of 1 or 0, and plural counters for counting the output of said signal level comparators.

Claim 11

An image processing apparatus according to claim 9, wherein said brightness sampling unit comprises a brightness signal converting means for composing brightness signals from the image signals.

Claim 12

An image processing apparatus according to claim 7, wherein said established nonlinear relation is between said histogram computation processing means output and said plurality of nonlinear gamma curves.

Claim 13

An image processing apparatus according to claim 1, wherein the nonlinear gamma correction of image signals is performed by a sigmoid curve to show nonlinearity.

Claim 14

An image processing apparatus according to claim 2, wherein the learning type gamma selector means further comprises a plurality of third recognition units, wherein each of the plurality of tree branch structures corresponds to one of a plurality of split picture blocks, wherein said network has five layers and said plurality of third recognition units are situated on the fifth layer and said plurality of second recognition units are situated on the fourth layer, each of said third recognition units comprises a summation computing unit for obtaining a total sum of input signals and a threshold processing unit for threshold processing said total sum of input signals and outputting results, each of said plurality of third recognition units is interlinked with each of the plurality of second recognition units, a plurality of groups of three brightness values are output by the image signal sampling means, said three brightness values corresponding to maximum, medium and minimum levels of brightness, respectively, wherein the number of groups of brightness values corresponds to the number of tree branch structures, wherein said plurality of groups of brightness values are inputted respectively to the signal inputting unit of said plurality of first recognition units situated on the first, the second and the third layers of the plurality of tree branch structures, and the selected non linear gamma curve to be used for correction is inputted as the teaching signal.

Claim 15

An image processing apparatus according to claim 2, wherein the learning type gamma selector means further comprises a plurality of third recognition units, each of the plurality of third recognition units is interlinked with each of the plurality of second recognition units, each of the plurality of tree branch structures corresponds to each of a plurality of image signal levels, each of the plurality of third recognition units comprises a summation computing unit for obtaining a total sum of input signals and a threshold processing unit for threshold processing the total sum of input signals and outputting results, respective frequencies for each of the plurality of image signal levels are output by the image signal sampling means, said respective frequencies are inputted to the signal inputting units of the first recognition units situated in the tree branch structures corresponding to each respective image signal level, and the selected nonlinear gamma curve to be used for correction is inputted as the teaching signal.