US5787205A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Motion vector detecting device
Assignee: Canon Inc
•
Filing: 1997-09-18
•
Granted: 1998-07-28
•
Expires: 2017-09-18
Abstract
Abstract The present invention provides a motion vector detecting device or method by which dependable motion vectors are detected. The device, according to one aspect of the invention, searches for a similar block in a previous image to a subject block in both a first area expanding from a corresponding position to the subject block and a second area expanding from a position indicated by a predictive vector and detecting a motion vector according to a result of the searching. Because there is a fair possibility of occurring the similar block in both the first and second areas.
Inventors
- 👤 Yasuji Hirabayashi
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-10 02:10
CPC Classifications
H; H04; H04N; H04N19/00; H04N19/50; H04N19/503; H04N19/51; H04N19/56
Claims Analysis
Total Claims: 30
Claim 1
Independent Claim
151 words
A motion vector detecting device comprising: input means for sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for predicting a motion vector for a subject block in a current image based on a registered vector in a previous image; detecting means for detecting a motion vector for the subject block by searching for a block similar to the subject block among selected blocks in the previous image, the selected blocks being only in a first area of the previous image which extends out from a position corresponding to the subject block in the previous image, the first area corresponding to a still area of the previous image, and in a second area of the previous image which extends out from a position indicated by the registered vector, the second area corresponding to a motion area of the previous image.
Claim 2
Dependent Claim
151 words
A motion vector detecting device comprising: input means for sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for predicting a motion vector for a subject block in a current image based on a registered vector in a previous image; detecting means for detecting a motion vector for the subject block by searching for a block similar to the subject block among selected blocks in the previous image, the selected blocks being only in a first area of the previous image which extends out from a position corresponding to the subject block in the previous image, the first area corresponding to a still area of the previous image, and in a second area of the previous image which extends out from a position indicated by the registered vector, the second area corresponding to a motion area of the previous image.
Claim 3
Independent Claim
156 words
A method for detecting motion vectors in a current images comprising the steps of: sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; storing registered motion vectors for blocks of a previous one of the plurality of images; and detecting a motion vector for a subject block in a current image by searching for a block similar to the subject block among selected blocks of the previous image, the selected blocks being only in a first area and in a second area of the previous image; wherein the first area of the previous image corresponds to a position of the subject block in the current image, the first area corresponding to a still area of the image; and wherein the second area corresponds to a position indicated by a registered vector corresponding to the subject block, the second area corresponding to a motion area of the image.
Claim 4
Independent Claim
156 words
A method for detecting motion vectors in a current images comprising the steps of: sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; storing registered motion vectors for blocks of a previous one of the plurality of images; and detecting a motion vector for a subject block in a current image by searching for a block similar to the subject block among selected blocks of the previous image, the selected blocks being only in a first area and in a second area of the previous image; wherein the first area of the previous image corresponds to a position of the subject block in the current image, the first area corresponding to a still area of the image; and wherein the second area corresponds to a position indicated by a registered vector corresponding to the subject block, the second area corresponding to a motion area of the image.
Claim 5
Dependent Claim
23 words
A device according to claim 2, wherein said predicting means disregards a previous motion vector which has a value within a predetermined range.
Claim 6
Dependent Claim
25 words
A device according to claim 3, wherein said predetermined range is determined to correspond to an area within the first area of the previous image.
Claim 7
Dependent Claim
42 words
A device according to claim 3, wherein said predicting means compensates for the registered vector for the subject block by using a previous registered vector when the value of the registered vector for the subject block is recognized as an invalid value.
Claim 8
Dependent Claim
29 words
A device according to claim 5, wherein said predicting means maintains a value of the previous registered vector when the motion vector has a value within the predetermined range.
Claim 9
Dependent Claim
32 words
A device according to claim 1, further comprising previous image preparing means which outputs pixels of the previous image in an order which is controlled on the basis of the motion vector.
Claim 10
Dependent Claim
27 words
A device according to claim 7, further comprising encoding means for encoding difference values between the pixels of the previous image and pixels of the present image.
Claim 11
Dependent Claim
154 words
A motion vector detecting method comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting a motion vector for a subject block in a current image based on a registered vector in a previous image; searching for a block similar to the subject block among selected blocks in a previous image, the selected blocks being only in a first area of the previous image which extends out from a position corresponding to the subject block in the previous image, the first area corresponding to a still area of the previous image, and in a second area of the previous image which extends out from a position indicated by the registered vector, the second area corresponding to a motion area of the previous image; and detecting a motion vector for the subject block according to a result of said step of searching.
Claim 12
Dependent Claim
154 words
A motion vector detecting method comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting a motion vector for a subject block in a current image based on a registered vector in a previous image; searching for a block similar to the subject block among selected blocks in a previous image, the selected blocks being only in a first area of the previous image which extends out from a position corresponding to the subject block in the previous image, the first area corresponding to a still area of the previous image, and in a second area of the previous image which extends out from a position indicated by the registered vector, the second area corresponding to a motion area of the previous image; and detecting a motion vector for the subject block according to a result of said step of searching.
Claim 13
Dependent Claim
132 words
A motion vector detecting device, comprising: input means for successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block; and detecting means for detecting a motion vector for the subject block by using a previous one of the plurality of images, the current image, and the predictive vector; wherein the predicting means produces the predictive vector by using a motion vector for a previous block in the previous image; and wherein, when producing the predictive vector, the predicting means ignores motion vectors for the previous block which have a value within a predetermined-range.
Claim 14
Dependent Claim
132 words
A motion vector detecting device, comprising: input means for successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block; and detecting means for detecting a motion vector for the subject block by using a previous one of the plurality of images, the current image, and the predictive vector; wherein the predicting means produces the predictive vector by using a motion vector for a previous block in the previous image; and wherein, when producing the predictive vector, the predicting means ignores motion vectors for the previous block which have a value within a predetermined-range.
Claim 15
Dependent Claim
43 words
A device according to claim 10, wherein said predicting means makes up for the predictive vector for the subject block by using a previous predictive vector when the value of the predictive vector for the subject block is recognized as an invalid value.
Claim 16
Dependent Claim
43 words
A device according to claim 10, wherein said predicting means makes up for the predictive vector for the subject block by using a previous predictive vector when the value of the predictive vector for the subject block is recognized as an invalid value.
Claim 17
Dependent Claim
28 words
A device according to claim 10, wherein said detecting means searches for a similar block in the previous image to the subject block to decide the motion vector.
Claim 18
Dependent Claim
44 words
A device according to claim 13, wherein said detecting means searches for the similar block in both a first area extends out to a position corresponding to the subject block and a second area extends out to a position indicated by the predictive vector.
Claim 19
Dependent Claim
18 words
A device according to claim 14, wherein said predetermined range corresponds to the extent of the first area.
Claim 20
Dependent Claim
44 words
A device according to claim 10, further including previous image preparing means which outputs pixels of the previous image in an order which is controlled on the basis of the motion vector even if the motion vector has a value within the predetermined range.
Claim 21
Dependent Claim
27 words
A device according to claim 16, further comprising encoding means for encoding difference values between the pixels of the previous image and pixels of the present image.
Claim 22
Dependent Claim
127 words
A motion vector detecting method, comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block; and detecting a motion vector for the subject block by using a previous one of the Plurality of images, the current image, and the predictive vector; wherein the producing step produces the predictive vector by using a motion sector for a previous block in the previous image; and wherein, when producing the predictive vector, the producing step ignores motion vectors for the previous block which have a value within a predetermined range.
Claim 23
Dependent Claim
127 words
A motion vector detecting method, comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block; and detecting a motion vector for the subject block by using a previous one of the Plurality of images, the current image, and the predictive vector; wherein the producing step produces the predictive vector by using a motion sector for a previous block in the previous image; and wherein, when producing the predictive vector, the producing step ignores motion vectors for the previous block which have a value within a predetermined range.
Claim 24
Dependent Claim
166 words
A motion vector detecting device comprising: input means for successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block and being used to define a search area for detecting a motion vector in a motion area of a previous image; and detecting means for detecting a motion vector for the subject block by using a previous one of the plurality of images, the current images and the predictive vector, wherein said predicting means replaces a value of the predictive vector for the subject block with a value of a previous predictive vector, the previous predictive vector having been produced for a block in the previous image, in a case where the value of the predictive vector for the subject block is recognized as being an invalid value.
Claim 25
Dependent Claim
166 words
A motion vector detecting device comprising: input means for successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; predicting means for producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block and being used to define a search area for detecting a motion vector in a motion area of a previous image; and detecting means for detecting a motion vector for the subject block by using a previous one of the plurality of images, the current images and the predictive vector, wherein said predicting means replaces a value of the predictive vector for the subject block with a value of a previous predictive vector, the previous predictive vector having been produced for a block in the previous image, in a case where the value of the predictive vector for the subject block is recognized as being an invalid value.
Claim 26
Dependent Claim
42 words
A device according to claim 18, wherein said predicting means replaces the invalid value of the predictive vector for the subject block with a value of the preceding predictive vector for a block being at the same position to the subject block.
Claim 27
Dependent Claim
159 words
A motion vector detecting method, comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block and being used to define a search area for detecting a motion vector in a motion area of a previous image; detecting a motion vector for the subject block by using a previous one of the plurality of images, the current images and the predictive vector, wherein the producing step replaces a value of the predictive vector for the subject block with a value of a previous predictive vector, the previous predictive vector having been produced for a block in the previous image, in a case where the value of the predictive vector for the subject block is recognized as being an invalid value.
Claim 28
Dependent Claim
159 words
A motion vector detecting method, comprising the steps of: successively inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; producing a predictive vector for a subject block in a current one of the plurality of images, the predictive vector predicting a movement direction of the subject block and being used to define a search area for detecting a motion vector in a motion area of a previous image; detecting a motion vector for the subject block by using a previous one of the plurality of images, the current images and the predictive vector, wherein the producing step replaces a value of the predictive vector for the subject block with a value of a previous predictive vector, the previous predictive vector having been produced for a block in the previous image, in a case where the value of the predictive vector for the subject block is recognized as being an invalid value.
Claim 29
Independent Claim
169 words
A device for detecting motion vectors in a current images comprising: input means for sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; storing means for storing registered motion vectors for blocks of a previous one of the plurality of images; and detecting means for detesting a motion vector for a subject block in a current one of the plurality of images by searching for a block similar to the subject block among selected blocks of the previous image, the selected blocks being only in a first area and in a second area of the previous image; wherein the first area of the previous image corresponds to a position of the subject block in the current image, the first area corresponding to a still area of the previous image; and wherein the second area corresponds to a position indicated by a registered vector corresponding to the subject block, the second area corresponding to a motion area of the previous image.
Claim 30
Independent Claim
169 words
A device for detecting motion vectors in a current images comprising: input means for sequentially inputting a plurality of images, each image being divided into blocks comprised of a plurality of pixels; storing means for storing registered motion vectors for blocks of a previous one of the plurality of images; and detecting means for detesting a motion vector for a subject block in a current one of the plurality of images by searching for a block similar to the subject block among selected blocks of the previous image, the selected blocks being only in a first area and in a second area of the previous image; wherein the first area of the previous image corresponds to a position of the subject block in the current image, the first area corresponding to a still area of the previous image; and wherein the second area corresponds to a position indicated by a registered vector corresponding to the subject block, the second area corresponding to a motion area of the previous image.
References Cited (Backward Citations)
US4562468A
Cited
US5162923A
Cited
US5031039A
Cited
US5089887A
Cited
EP0360698A1
Cited
US5086487A
Cited
JPH02294180A
Cited
US5043808A
Cited
US5150209A
Cited
US5227877A
Cited
JPH04127690A
Cited
WO1992019068A1
Cited
US5258836A
Cited
US5210605A
Cited
EP0518314A2
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-07-27 | 2002-01-26 | 2002-07-28 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-07-27 | 2006-01-26 | 2006-07-28 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-07-27 | 2010-01-26 | 2010-07-28 | $7700.00 | Window Closed / Lapsed |