EP0710004A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Image processing apparatus
Assignee: Sharp Corp
•
Filing: 1995-10-11
•
Granted: —
•
Expires: 2015-10-11
Abstract
Abstract An image processing apparatus inputs an image signal obtained by scanning a document in which a character region, a photographic region and a dot region are mixed and stores image data in a local block composed of a target picture element and plural picture elements surrounding the target picture element. The image processing apparatus computes first and second feature parameters P₀ and P₁ representing the features of each region based on the image data in the local block, and inputs the resulting first and second feature parameters P₀ and P₁ to an identification circuit adopting a neural network. The identification circuit outputs the region identification information O₀ and O₁ of the target picture element, and the filter processing circuit performs a spatial filtering process according to the region identification information O₀ and O₁. As described, in the image processing apparatus, since a multi-dimensional identification process is performed using the neural circuit which receives inputs of plural feature parameters, an image identification with very high precision is permitted, and an optimum spatial-filtering process can be applied according to the feature of each picture element, thereby achieving a high quality image.
Inventors
- 👤 Yasushi Adachi
- 👤 Yoshiyuki Nakai
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 22:33
CPC Classifications
H; H04; H04N; H04N1/00; H04N1/40; H04N1/40062
Claims Analysis
Total Claims: 31
Claim 1
Independent Claim
184 words
An image processing apparatus, comprising: a block memory for storing image data in a local block composed of a target picture element and plural picture elements surrounding said target picture element, said target picture element being one of picture element data of an image signal; feature parameter computing means for computing plural feature parameters based on the image data in the local block stored in said block memory; identification processing means for outputting region identification information regarding a region where the target picture element in the local block is located based on inputs representing respective feature parameters computed by said feature parameter computing means; and filtering means for selecting a filter among spatial-filtering-use filters of various types based on an output from said identification processing means and applying a spatial-filtering process to the target picture element; wherein said identification processing means divides a plane or a multi-dimensional space having axes of the plural feature parameters into a character region, a photographic region and a dot region by non-linear boundaries and outputs respective region identification information by receiving each input of the plural feature parameters.
Claim 2
Dependent Claim
42 words
The image processing apparatus as set forth in claim 1, wherein: said identification processing means is composed of a neural network which learned well beforehand, said neural network receiving each input of the plural feature parameters and outputting said region identification information.
Claim 3
Dependent Claim
51 words
The image processing apparatus as set forth in claim 1, wherein: said identification processing means is composed of a look up table describing an input-output correlation of a neural network which learned well beforehand, said neural network receiving each input of the plural feature parameters and outputting said region identification information.
Claim 4
Dependent Claim
66 words
The image processing apparatus as set forth in claim 1, wherein: region identification information outputted from said identification processing means is an identification signal for identifying that the target picture element in the local block is located in which one of the character region, the photographic region and the dot region, and each filter has a filter factor which is predetermined according to each spatial-filtering process.
Claim 5
Dependent Claim
96 words
The image processing apparatus as set forth in claim 4, wherein: said filtering means performs a spatial-filtering process using an enhancement filter when it is determined that the target picture element is located in the character region based on the region identification information, while performs a spatial-filtering process using a smoothing filter when it is determined that the target picture element is located in the dot region, and performs a spatial-filtering process for outputting the target picture element without processing it when it is determined that the target picture element is located in the photographic region.
Claim 6
Dependent Claim
53 words
The image processing apparatus as set forth in claim 1, wherein: said feature parameter computing means includes a signal level difference computing means for outputting a difference between a maximum signal level and a minimum signal level in the local block to said identification processing means as one of the plural feature parameters.
Claim 7
Dependent Claim
118 words
The image processing apparatus as set forth in claim 1, wherein said feature parameter computing means includes: main-scanning direction difference sum computing means for computing a sum of differences in signal level between picture elements in a local block which are in sequence in the main-scanning direction; sub-scanning direction difference sum computing means for computing a sum of differences in signal level between picture elements in the local block which are in sequence in the sub-scanning direction; and minimum value computing means for outputting a minimum value of computation results by said main-scanning direction difference sum computing means and the sub-scanning direction difference sum computing means to said identification processing means as one of the plural feature parameters.
Claim 8
Dependent Claim
127 words
The image processing apparatus as set forth in claim 1, wherein said feature parameter computing means includes: binary means for converting signal levels of the target picture element and the plural picture elements surrounding the target picture element in the local block to one of two numerals; main-scanning direction change point counter means for counting a number of changes in signal level between two values in the main-scanning direction; sub-scanning direction change point counter means for counting a number of changes in signal level between two values in the sub-scanning direction; and minimum value computing means for outputting a minimum value of count results by said main-scanning direction change point counter means and said sub-scanning direction change point counter means as one of the plural feature parameters.
Claim 9
Dependent Claim
129 words
The image processing apparatus as set forth in claim 1, wherein said feature parameter computing means includes: binary means for converting signal levels of the target picture element and the plural picture elements surrounding the targets picture element in the local block to one of two numerals; main-scanning direction run-length counter means for computing a run-length in the main-scanning direction of a signal level converted to one of the two numerals; sub-scanning direction run-length counter means for computing a run-length in the sub-scanning direction of a signal level converted to one of the two numerals; and maximum value computing means for outputting a maximum value of count results by said main-scanning direction run-length counter means and said sub-scanning direction run-length counter means as one of the plural feature parameters.
Claim 10
Dependent Claim
151 words
The image processing apparatus as set forth in claim 1, wherein said feature parameter computing means includes: main-scanning direction feature parameter computing means for computing a feature parameter in the main-scanning direction in the local block; sub-scanning direction feature parameter computing means for computing a feature parameter in the sub-scanning direction in the local block; block feature parameter computing means for computing a block feature parameter of the local block based on feature parameters computed respectively by said main-scanning direction feature parameter computing means and said sub-scanning direction feature parameter computing means; and a weighted mean accumulating section for computing a weighted average of said block feature parameter and a feature parameter of a lastly processed local block, and the weighted mean computed by said weighted mean accumulating section is outputted to said identification processing means as one of the plural feature parameters of the local block to be currently processed.
Claim 11
Dependent Claim
67 words
The image processing apparatus as set forth in claim 10, wherein: said weighted mean accumulating section computes a weighted mean P₁ from formula (1) P 1 = (αP 1a + βP 1b )/(α + β) (α > β) wherein P 1a is a block feature parameter computed by said block feature parameter computing means, and P 1b is the feature parameter in the lastly processed local block.
Claim 12
Dependent Claim
147 words
The image processing apparatus as set forth in claim 10, wherein: said main-scanning direction feature parameter computing means is composed of main-scanning direction difference sum computing means for computing a sum of differences in signal level between two picture elements in the local block which are in sequence in a main-scanning direction; said sub-scanning direction feature parameter computing means is composed of sub-scanning direction difference sum computing means for computing a sum of differences in signal level between two picture elements in the local block which are in sequence in a sub-scanning direction; and said block feature parameter computing means is composed of minimum value computing means for computing a minimum value of computation results by said main-scanning direction difference sum computing means and said sub-scanning direction difference sum computing means and outputting the minimum value to said weighted mean accumulating section as the block feature parameter.
Claim 13
Dependent Claim
148 words
The image processing apparatus as set forth in claim 10, wherein: said feature parameter computing means further includes binary means for converting signal levels of the target picture element and the plural picture elements in the local block to one of two numerals, said main-scanning direction feature parameter computing means is composed of a main-scanning direction change point counter means for counting a number of changes in signal level between two values in the main-scanning direction, said sub-scanning direction feature parameter computing means is composed of a sub-scanning direction change point counter means for counting a number of changes in signal level between two values in the sub-scanning direction, and said block feature parameter computing means is composed of minimum value computing means for computing a minimum value resulting from respective computations by said main-scanning direction change point counter means and said sub-scanning direction change point counter means.
Claim 14
Dependent Claim
191 words
The image processing apparatus as set forth in claim 10, wherein: said feature parameter computing means further includes binary means for converting signal levels in the target picture element and the plural picture elements in the local block to one of two numerals, said main-scanning direction feature parameter computing means is composed of main-scanning direction run-length counter means for counting a run-length in the main-scanning direction of signal levels of the target picture element and the plural picture elements surrounding the target picture element converted to one of the two numerals, said sub-scanning direction feature parameter computing means is composed of sub-scanning direction run-length counter means for counting a run-length in the sub-scanning direction of signal levels of the target picture element and the plural picture elements surrounding the target picture element converted to one of the two numerals, and said block feature parameter computing means is composed of maximum value computing means for computing a maximum value of computation results by said main-scanning direction run-length counter means and said sub-scanning direction run-length counter means and outputting the maximum value to said weighted mean accumulating section as the block feature parameter.
Claim 15
Dependent Claim
81 words
The image processing apparatus as set forth in claim 1, further comprising: memory means for storing contents of spatial-filtering processes applied to the plural picture elements surrounding the target picture element, wherein said filtering means determines a content of a spatial-filtering process to be applied to the target picture element based on not only region identification information from said identification processing means but also the contents of the spatial-filtering processes applied to the plural picture elements surrounding the target picture element.
Claim 16
Dependent Claim
43 words
The image processing apparatus as set forth in claim 15, wherein: said memory means includes an application number memory section for storing an accumulated number of applying each spatial-filtering process to a predetermined number of picture elements before processing the target picture element.
Claim 17
Dependent Claim
176 words
An image processing apparatus, comprising: a block memory for storing image data in a local block composed of a target picture element and plural picture elements surrounding the target picture element, the target picture element being one picture element data of an image signal; feature parameter computing means for computing respective plural feature parameters from the image data in the local block stored in said block memory; identification processing means for outputting region identification information representing respective likelihoods that the target picture element is located in a character region, a photographic region and a dot region; and filtering means for selecting a filter among spatial-filtering-use filters of various types based on an output from said identification processing means and applying a spatial-filtering process to the target picture element, wherein said identification processing means divides a plane or a multi-dimensional space having axes of the plural feature parameters into the character region, the photographic region and the dot region by non-linear boundaries and outputs respective region identification information according to each input of the plural feature parameters.
Claim 18
Dependent Claim
42 words
The image processing apparatus as set forth in claim 17, wherein: said identification processing means is composed of a neural network which learned well beforehand, said neural network receiving each input of the plural feature parameters and outputting the region identification information.
Claim 19
Dependent Claim
51 words
The image processing apparatus as set forth in claim 17, wherein: said identification processing means is composed of a look up table describing an input-output correlation of a neural network which learned well beforehand, said neural network receiving an input of the plural feature parameters and outputting said region identification information.
Claim 20
Dependent Claim
97 words
The image processing apparatus as set forth in claim 17, wherein: said filtering means includes comparison means for making a comparison among said respective likelihoods represented by said region identification information outputted from said identification processing means, and when said comparison means determines a region which shows a highest likelihood, said filtering means selects a filter for a spatial-filtering process which is suited for the region, while when said comparison means determines that there is no significant difference in the likelihoods, said filtering means performs a spatial-filtering process for outputting the target picture element without processing it.
Claim 21
Dependent Claim
81 words
The image processing apparatus as set forth in claim 17, wherein: said filtering means includes comparison means for comparing among said respective likelihoods represented by said region identification information outputted from said identification processing means, when said comparison means determines a region which shows a highest likelihood, a spatial-filtering-use filter for the region is selected, while when said comparison means determines that there is no significant difference in said likelihoods, a spatial-filtering process in combination of various spatial-filtering processes is performed.
Claim 22
Dependent Claim
78 words
The image processing apparatus as set forth in claim 17, further comprising: memory means for storing a content of a spatial-filtering process applied to a picture element prior to the target picture element; and said filtering means includes correction means for correcting the region identification information from said identification processing means according to the content of the spatial-filtering process stored in said memory means, that is applied to the picture element directly before processing the target picture element.
Claim 23
Dependent Claim
66 words
The image processing apparatus as set forth in claim 17, wherein: said region identification information from said identification processing means is an identification signal for identifying that the target picture element in the local block is located in which one of the character region, the photographic region and the dot region, and each filter has a filter factor which is predetermined according to each spatial-filtering process.
Claim 24
Dependent Claim
150 words
The image processing apparatus as set forth in claim 17, wherein said feature parameter computing means includes: main-scanning direction feature parameter computing means for computing a feature parameter in the main-scanning direction in the local block; sub-scanning direction feature parameter computing means for computing a feature parameter in the sub-scanning direction in the local block; block feature parameter computing means for computing a block feature parameter of the local block based on respective feature parameters computed by said main-scanning direction feature parameter computing means and said sub-scanning direction feature parameter computing means; and weighted mean accumulating section for computing a weighted mean of said block feature parameter and the feature parameter of a lastly processed local block, and the weighted mean computed by said weighted mean accumulating section is outputted to said identification processing means as one of the plural feature parameters of the local block to be currently processed.
Claim 25
Dependent Claim
82 words
The image processing apparatus as set forth in claim 17, further comprising: memory means for storing contents of spatial-filtering processes applied to the plural picture elements surrounding the target picture element, wherein said filtering means determines a content of a spatial-filtering process to be applied to the target picture element based on not only the region identification information from said identification processing means but also the contents of the spatial-filtering processes applied to the plural picture elements surrounding the target picture element.
Claim 26
Dependent Claim
45 words
The image processing apparatus as set forth in claim 25, wherein: said memory means includes an application number memory section for storing an accumulated number of applying each spatial-filtering process applied to a predetermined number of picture elements prior to processing the target picture element.
Claim 27
Dependent Claim
152 words
An image processing apparatus, comprising: a block memory for storing image data in a local block composed of a target picture element and plural picture elements surrounding the target picture element, the target picture element being one image data of an image signal; feature parameter computing means for computing plural feature parameters based on the image data in the local block stored in said block memory; identification processing means for outputting region identification information representing respective likelihoods that the target picture element is located in a character region, a photographic region and a dot region; and filtering means for applying a spatial-filtering process to the target picture element using a filter; wherein filtering means includes filter factor computing means for computing a filter factor suited for the target picture element based on the region identification information from said identification means, said filter factor computing means setting the filter factor in the filter.
Claim 28
Dependent Claim
70 words
The image processing apparatus as set forth in claim 27, wherein: said identification processing means divides a plane or a multi-dimensional space having axes of the plural feature parameters into the character region, the photographic region and the dot region by non-linear boundaries, said identification processing means being composed of a neural network which learned well beforehand for outputting each likelihood by receiving each input of the plural feature parameters.
Claim 29
Dependent Claim
81 words
The image processing apparatus as set forth in claim 27, wherein: said identification processing means divides a plane or a multi-dimensional space having axes of the plural feature parameters into the character region, the photographic region and the dot region by non-linear boundaries, said identification processing means being composed of a look up table describing an input-output correlation of a neural network which learned well beforehand, said numeral network outputting each likelihood by receiving each input of the plural feature parameters.
Claim 30
Dependent Claim
53 words
The image processing apparatus as set forth in claim 27, wherein: said filter factor computing means outputs the region identification information in a form of numeral representing each likelihood, and said filter factor computing means computes a filter factor for use in various spatial-filtering processes as a function of the region identification information.
Claim 31
Dependent Claim
99 words
The image processing apparatus as set forth in claim 27, further comprising: filter factor memory means for storing therein a filter factor used in processing a picture element directly before processing the target picture element, wherein said identification processing means outputs the region identification information in a form of numeral representing each likelihood, and said filter factor computing means outputs a filter factor for use in each spatial-filtering process as a function between the region identification information indicating the region in which the target picture element is located and the filter factors stored in said filter factor memory means.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.