Claims Analysis
Total Claims: 14
Claim 1
Independent Claim
111 words
An image processing apparatus comprising an image signal sampling unit wherein plural kinds of picture information are sampled from the image signals of one picture frame, a non linear gamma correction unit wherein the gamma values of said picture signals are corrected by utilizing nonlinear gamma curves, and a learning type gamma selector unit wherein a nonlinear relation between the picture information of image signals sampled by said image signal sampling unit and said nonlinear gamma curves to be used for correction is established by learning and said picture information is judged according to the results of said learning and a nonlinear gamma curve to be used for correction is selected.
Claim 2
Dependent Claim
180 words
An image processing apparatus according to Claim 1, consisting of a combination of a multiple number of a recognition unit comprising a signal inputting unit where in the picture information of image signals is inputted, a quantitizer wherein image signals are quantitized according to the input signals from said signal inputting unit, a single or multiple number of channel input terminal, a single or multiple number of channel output terminal, and a channel selector unit wherein channels are selected according to the output of said quantitizer, to form a single or multiple number of tree branch structure which comprises plural layers and forms networks, and the recognition unit situated on the front layers of said networks output layers comprising a signal input unit for inputting a teaching signal, a single or multiple number of channel input terminal, a single or multiple number of channel output terminal, and a channel learning unit wherein the strength of linkage between said channel input terminal and the channel output terminal appointed by the teaching input value inputted from said channel input unit is varied.
Claim 3
Dependent Claim
42 words
An image processing apparatus according to Claim 1, wherein an image signal sampling unit comprises a picture splitting unit for splitting image signals into plural picture blocks and a characteristic signal sampling unit for sampling image signals from said split picture blocks.
Claim 4
Dependent Claim
50 words
An image processing apparatus according to Claim 3, wherein a characteristic signal sampling unit comprises a brightness sampling unit for sampling brightness signals from said split picture blocks, and a brightness computation unit for computing the maximum, the medium and the minimum brightness values within each of said picture blocks.
Claim 5
Dependent Claim
30 words
An image processing apparatus according to Claim 4, wherein a brightness sampling unit comprises a brightness signal converting means for composing brightness signals from the R, G and B signals.
Claim 6
Dependent Claim
73 words
An image processing apparatus according to Claim 1, wherein a nonlinear relation between the image signals sampled for each picture block by the image signal sampling unit and the nonlinear gamma curves to be used for correction is established by learning and said image signals are judged according to the results of said learning for selection of a nonlinear gamma curve to be used for correction at the learning type gamma selector unit.
Claim 7
Dependent Claim
45 words
An image processing apparatus according to Claim 1, wherein a image signal sampling unit comprises a histogram computation processing means for inputting and processing the image signals of one picture frame, and obtaining the frequency for each of plural signal levels as the picture information.
Claim 8
Dependent Claim
61 words
An image processing apparatus according to Claim 7, wherein a histogram computation processing means comprises plural signal level comparators for comparing the levels of picture information with a level having a certain spread in its value and outputting the comparison results with a digit number of 1 or 0, and plural counters for counting the output of said signal level comparators.
Claim 9
Dependent Claim
73 words
An image processing apparatus according to Claim 7, wherein a histogram computation processing means comprises a brightness sampling unit for sampling brightness signals, plural signal level comparators for comparing the brightness signals sampled through said brightness sampling unit with a level having a certain spread in its value and outputting the comparison results with a digit number of 1 or 0, and plural counters for counting the output of said signal level comparators.
Claim 10
Dependent Claim
30 words
An image processing apparatus according to Claim 9, wherein a brightness sampling unit comprises a brightness signal converting means for composing brightness signals from the R, G, and B signals.
Claim 11
Dependent Claim
77 words
An image processing apparatus according to Claim 1, wherein a nonlinear relation between the frequency for each of plural image signal levels sampled by the image signal sampling unit and the nonlinear gamma curves to be used for correction is established by learning and the frequencies of said signal levels are judged according to the results of said learning for selecting a nonlinear gamma curve to be used for correction at the learning type gamma selector unit.
Claim 12
Dependent Claim
23 words
An image processing apparatus according to Claim 1, wherein the gamma of image signals is corrected by a sigmoid curve to show nonlinearity.
Claim 13
Dependent Claim
202 words
An image processing apparatus according to Claim 2, wherein a learning type gamma sampling unit comprises a combination of plural recognition units, each of plural tree branch structures corresponding in number to the number of split picture blocks forms a five layer network, the recognition unit of the fifth layer serving as an output layer comprises a summation computing unit for obtaining the total sum of input signals and a threshold processing unit for threshold processing said total sum of input signals and outputting the results thereof, each of said recognition units situated on the fifth layer is interlinked with the recognition units as carried by the fourth layer of each respective tree branch structure, three brightness values of the maximum, the medium and the minimum level sampled through the image signal sampling unit are inputted respectively to the signal input terminals of the recognition units situated on the first, the second and the third layers of the tree branch structures corresponding to each of the picture blocks, and a selected signal of the nonlinear gamma curve to be used for correction is inputted as a teaching signal to the signal input unit of said recognition unit situated on the fourth layer.
Claim 14
Dependent Claim
196 words
An image processing apparatus according to Claim 2, wherein a learning type gamma sampling unit comprises a combination of plural recognition units, a network is formed so as to have plural tree branch structures corresponding in number to the number of histogram signal levels obtained from plural image signal levels, the recognition units of the output layers comprises a summation computing unit for obtaining the total sum of input signals and a threshold processing unit for threshold processing the total sum of input signals and outputting the results thereof, each of the recognition units of said output layers is interlinked with each of the recognition units as carried by the front layers of the output layers situated in each tree branch structure, the image signals sampled through the image signal sampling unit are inputted to the signal input terminals of the recognition units situated in the tree branch structure for each respective histogram level of the picture information, and a selected signal of the nonlinear gamma curve to be used for correction is inputted as a teaching signal to the signal input unit of the recognition unit situated in the front layer of said output layer.