TOP > 外国特許検索 > IMAGE RECOVERY DEVICE, METHOD, AND PROGRAM

IMAGE RECOVERY DEVICE, METHOD, AND PROGRAM NEW コモンズ 新技術説明会

外国特許コード F180009450
整理番号 (T2016-023,S2017-0150-N0)
掲載日 2018年7月24日
出願国 世界知的所有権機関(WIPO)
国際出願番号 2017JP042106
国際公開番号 WO 2018097211
国際出願日 平成29年11月22日(2017.11.22)
国際公開日 平成30年5月31日(2018.5.31)
優先権データ
  • 特願2016-229777 (2016.11.28) JP
発明の名称 (英語) IMAGE RECOVERY DEVICE, METHOD, AND PROGRAM NEW コモンズ 新技術説明会
発明の概要(英語) The purpose of the present invention is to remove an irregular texture and background color included in image information. A determination unit (61) determines, on the basis of a reference value based on the pixel values of observation image information and the pixel values of pixels, whether the pixels are an irregular texture or background color. A state quantity calculation unit (62) calculates a state quantity of a state n of original image information using a state quantity of a state (n-1), a state transition from (n-1) to n, and a driving source. An observation quantity calculation unit (63) calculates an observation quantity representing a process of the original image information becoming the observation image information due to the addition of an irregular texture or background color, using the state quantity of the state n, the pixel value of a pixel determined to be an irregular texture or background color, and a non-linear term representing an influence on the process in accordance with the original image information. An estimation unit (64) estimates an optimum value of the state quantity of the state n as the original image information, using a Kalman algorithm equipped with a colored driving source and on the basis of the calculated state quantity and observation quantity.
特許請求の範囲(英語) [claim1]
1. Original image information in the at least one of the non-uniform texture and background color are mixed with respect to the observation image information, as a drive source of the original image and the image information based on the predicted state space model comprises using a method, the observation image information from the estimated original image information in the image reconstruction device, said observation included in the image information from the distribution of the pixel value of the pixel and the reference value is calculated, the observation included in the image information based on the value of each pixel of the, wherein the inert gas and the observation image information of each of the pixels of the background color and at least one of a uniform texture determination means for determining whether or not, in the original image and the time information of the state of the state quantity of an n n, (n-1) at the time the state of the state quantity of the (n-1), wherein said state n (n-1) state to state transition of, and the state of the state quantity (n-1) wherein n is not included in the state of the state amount of the driving source elements is calculated by using a state amount calculation means, wherein the inert gas and the original image information in the background to a uniform texture and at least one of the observation image is applied to the observed amount becomes information continuously in the, wherein the n state of the state amount, and the determination means determines coordinates of the background color and the non-uniform texture by at least one of the pixels determined by using the pixel value of the amount calculation means for calculating an observability, wherein the state quantity calculation unit calculates the state that has been calculated by the amount, calculated by the calculating means calculates the viewer and the surveying instrument based on the amount of the result of the observation, based on the predicted by said state space model, wherein the state of the state amount of the optimal value of n as original image information estimating means, the image decompression apparatus comprising.
[claim2]
2. Original image information in the at least one of the non-uniform texture and background color are mixed with respect to the observation image information, as a drive source of said original image information in the image based on the predicted state space model comprises using a method, the observation image information from the original image information in the image restoration device estimate, the observation included in the image information based on the pixel values of the pixels and the reference value, said observation included in the image information value and the pixel value of each pixel based on, wherein the inert gas and the observation image information each of the pixels of the background color of the at least one of a uniform texture and a determination means for determining whether or not, wherein n of the original image information in the time of the state quantity of an n state, the state at time (n-1) (n-1) of the state amount, wherein said state n (n-1) state to state transition of, and the state of the state quantity (n-1) wherein n is not included in the state of the state amount of the driving source elements is computed using the state amount calculation means, wherein the non-uniform texture information of the original background color and the at least one of the observed image is applied to the observed amount becomes information continuously in the, wherein the n state of the state amount, the determination means determines a uniform texture of the above-mentioned by at least one of the background color and a pixel value of a pixel is determined to be, wherein the step in accordance with the information of the original and the influence of the nonlinear term TSKGEL is calculated by using the observation quantity calculation device, which is calculated by the state quantity calculation unit calculates the state quantity, calculated by the calculating means calculates a viewing metric and based on the amount of the observation, based on the predicted by the state space model, wherein the state of the state amount of the optimal value of n is estimated as the original image information and the estimating means, the image decompression apparatus comprising.
[claim3]
3. Wherein the image information is observed, by reading a document having an image on both sides are obtained, in which the inert gas is a uniform texture, wherein the information of the original document surfaces having, on the other surface of the show and the show-through image is image information, wherein the metric calculation means calculates an image viewer, wherein the show-through image information to said non-linear term by multiplying, by adding the original image Im information, wherein the image restoring apparatus according to claim 2 calculated metrics viewer.
[claim4]
4. Wherein the metric calculation means calculates the image viewer, wherein the determining means determines the non-uniform texture and determined to be the pixel value of the pixel, wherein the image restoring apparatus according to claim 3 is used as a show-through image information.
[claim5]
5. Wherein the metric calculation means calculates the image viewer, wherein the back surface of the document obtained by reading the observation information of the back surface, wherein the image restoring apparatus according to claim 3 is used as a show-through image information.
[claim6]
6. In which the inert gas is a uniform texture, and the image information is included in the image showing the observation The wrinkles and fold, as well as the observation image information that has been applied to the background of the image indicated by the at least one of gradation, wherein the metric calculation means calculates the image viewer, wherein the non-uniform texture comprises the texture of the image information to said non-linear multiplication, by adding the original image information wherein the image restoring apparatus according to claim 2 calculated metrics viewer.
[claim7]
7. Said reference value is, the observation included in the image information and the distribution of the pixel value of the pixel to a value calculated from any one of claims 1 to claim 2 - claim 6 at the image decompression device.
[claim8]
8. Original image information in the at least one of the non-uniform texture and background color are mixed with respect to the observation image information, as a drive source of said original image information in the image based on the predicted state space model comprises using a method, the observation image information from the original image information in the image restoration method estimates, when the determination means, said observation included in the image information of a pixel value of the pixel and the reference value calculated from the distribution, wherein the observation included in the image information based on the value and the pixel value of each pixel, wherein each pixel of the observed image of the non-uniform texture information and determines whether or not the at least one of the background, when the state quantity calculation unit calculates, at the time of said original image information in the state of the state quantity of an n n, (n-1) at the time the state of the state quantity of the (n-1), wherein said state n (n-1) state to state transition of, and the state of the state quantity (n-1) that is not included in the state of the state amount of the n elements is calculated by using the driving power source, the measurement means and the viewer, wherein the non-uniform texture information of the original background color and the at least one of the observation image is applied as the information about the amount of observation represents processes, wherein the n state of the state amount, and the determining means determines the non-uniform texture and at least one of the background color of the pixel value of a pixel is determined using the calculated, estimated means, wherein the state quantity calculation unit calculates the state amount calculated by the, viewer and calculated by the calculating means calculates a metric based on the amount of observation, wherein the prediction method based on the state space model, wherein the optimal value of the n state of the state quantity is estimated as the original image information in the image reconstruction method.
[claim9]
9. Original image information in the at least one of the non-uniform texture and background color are mixed with respect to the observation image information, as a drive source of the original image and the image information based on the predicted state space model comprises using a method, the observation image information from the original image information in the image restoration method estimates, when the determination means, said observation included in the image information based on the pixel values of the pixels and the reference value, said observation included in the image information based on the value of each pixel of the, wherein each of the pixels of the observation image information of the non-uniform texture and determining whether or not the at least one of the background color, when the state quantity calculation unit calculates, at the time of the original image information in the state of the state quantity of an n n, (n-1) at the time the state of the state quantity of the (n-1), wherein said state n (n-1) state to state transition of, and the state of the state quantity (n-1) is not included in the state of the state amount of the n elements is calculated by using the driving power source, means for surveying the image viewer, wherein the non-uniform texture information of the original background color and the at least one of the observed image is applied to the observed amount becomes information continuously in the, wherein the state quantity of the n state, in which the inert gas by the determination means determines a uniform texture and at least one of the background color of the pixel value of a pixel determined to be, and in response to the original image information in the non-linear term represents the influence on the course calculated by using, when the estimation unit, which is calculated by the state quantity calculation unit calculates the state quantity, calculated by the calculating means calculates a viewing metric and based on the amount of the observation, wherein the prediction method based on the state space model, wherein the state of the state amount of the optimal value of n is estimated as the original image information in the image reconstruction method.
[claim10]
10. A computer, any of the claim 1 - claim 7 1 constituting the image decompression device to function as each means for causing the image decompressing program.
  • 出願人(英語)
  • ※2012年7月以前掲載分については米国以外のすべての指定国
  • TOKYO UNIVERSITY OF SCIENCE EDUCATIONAL FOUNDATION
  • TOKYO UNIV OF SCIENCE
  • 発明者(英語)
  • TANABE NARI
  • NATORI TAKAHIRO
  • FURUKAWA TOSHIHIRO
国際特許分類(IPC)
上記の特許・技術に関心のある方は、下記問合せ先にご相談下さい。

PAGE TOP

close
close
close
close
close
close