US4484349A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Parallel pipeline image processor
Assignee: ERIM International Inc
•
Filing: 1982-03-11
•
Granted: 1984-11-20
•
Expires: 2002-03-11
Abstract
Abstract A plurality of serial neighborhood transformation pipelines are provided for simultaneously operating on adjacent segments of a partitioned image matrix. Techniques are disclosed for bi-directionally transferring pixel data on the edges of adjoining segments of the image matrix between adjacent processors in a manner which minimizes the number of interconnections therebetween. In such manner a parallel pipeline image processing system can be implemented in integrated circuit form while keeping the number of pins for each stage in the pipeline to a minimum.
Inventors
- 👤 David L. McCubbrey
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-24 11:37
CPC Classifications
G; G06; G06F; G06F1/00; G06F1/22; G06V; G06V10/00; G06V10/20; G06V10/36
Claims Analysis
Total Claims: 14
Claim 1
Independent Claim
153 words
A method of analyzing a matrix of pixel values in a serial neighborhood processing system, said system being adapted for sequentially performing a neighborhood transformation on pixel values contained in an array of neighborhood extraction storage devices, said method comprising: partitioning the matrix into a plurality of segments; feeding contiguous segments to adjacent neighborhood transformation stages in a raster scan line format; sensing the relative position of each pixel in the lines fed to each stage; bi-directionally transferring pixels lying on the edges of each segment to adjacent stages over a single connection means before said pixels are loaded into the neighborhood extraction devices for processing; and then, loading neighboring pixel values into said neighborhood extraction devices for processing whereby each stage contains all of the neighboring pixels for performing said neighborhood transformations on the pixels fed to the stage even though some of the neighboring pixels were initially fed to adjacent stages.
Claim 2
Independent Claim
153 words
A method of analyzing a matrix of pixel values in a serial neighborhood processing system, said system being adapted for sequentially performing a neighborhood transformation on pixel values contained in an array of neighborhood extraction storage devices, said method comprising: partitioning the matrix into a plurality of segments; feeding contiguous segments to adjacent neighborhood transformation stages in a raster scan line format; sensing the relative position of each pixel in the lines fed to each stage; bi-directionally transferring pixels lying on the edges of each segment to adjacent stages over a single connection means before said pixels are loaded into the neighborhood extraction devices for processing; and then, loading neighboring pixel values into said neighborhood extraction devices for processing whereby each stage contains all of the neighboring pixels for performing said neighborhood transformations on the pixels fed to the stage even though some of the neighboring pixels were initially fed to adjacent stages.
Claim 3
Dependent Claim
36 words
The method of claim 1 wherein the feeding of each line of pixels to outer stages in the system is delayed by a suitable amount so as to prevent simultaneous transfer of pixels in a stage.
Claim 4
Dependent Claim
67 words
In a parallel partitioned serial neighborhood processing system having a plurality of pipelines of serial neighborhood transformation stages, adjacent pipelines being adapted to perform neighborhood transformations on contiguous segments of pixel data fed to each pipeline, wherein the improvement comprises: a single connection for pixel transfer between adjacent stages in the pipelines for bi-directionally transferring pixels lying on the edges of the segments fed to adjacent stages.
Claim 5
Dependent Claim
67 words
In a parallel partitioned serial neighborhood processing system having a plurality of pipelines of serial neighborhood transformation stages, adjacent pipelines being adapted to perform neighborhood transformations on contiguous segments of pixel data fed to each pipeline, wherein the improvement comprises: a single connection for pixel transfer between adjacent stages in the pipelines for bi-directionally transferring pixels lying on the edges of the segments fed to adjacent stages.
Claim 6
Dependent Claim
114 words
The improvement of claim 3 wherein each stage further comprises: transfer control means including multiplexer means for receiving pixel data from the matrix segment supplied to its pipeline and from said pixel transfer connections from adjacent stages in other pipelines, the output of said multiplexer means being coupled to neighborhood extraction means for temporarily storing a neighborhood of pixel values for processing and also being coupled to said pixel transfer connections; and means for sensing the relative position of the pixels in each segment line fed to the stage, operative to control the loading of the neighborhood extraction means through the multiplexer and transfer of pixels to adjacent stages through said pixel transfer connections.
Claim 7
Dependent Claim
66 words
A serial neighborhood transformation stage comprising: neighborhood extraction means for sequentially accessing a neighborhood of pixels in a matrix; transformation logic means for transforming each pixel in a matrix as a function of its value and surrounding pixel values contained in the neighborhood extraction means; and transfer control means up stream of said neighborhood extraction means, operative to receive and transfer selected pixels between other stages.
Claim 8
Dependent Claim
66 words
A serial neighborhood transformation stage comprising: neighborhood extraction means for sequentially accessing a neighborhood of pixels in a matrix; transformation logic means for transforming each pixel in a matrix as a function of its value and surrounding pixel values contained in the neighborhood extraction means; and transfer control means up stream of said neighborhood extraction means, operative to receive and transfer selected pixels between other stages.
Claim 9
Dependent Claim
150 words
A parallel partitioned serial neighborhood processing system comprising: means for partitioning an image matrix into a plurality of contiguous segments; means for feeding contiguous segments to adjacent serial neighborhood transformation stages in a raster scan line format; said serial neighborhood transformation stages including neighborhood extraction means for sequentially accessing a central pixel and its surrounding pixels in the matrix, neighborhood transformation logic means for transforming the value of the central pixel as a function of its value and the values of its surrounding pixels contained in the neighborhood extraction means; a single connection for pixel transfer between adjacent stages; means for sensing the relative position of each pixel in the lines fed to each stage; and transfer control means for bi-directionally transferring pixels lying on the edges of each segment to adjacent stages over said pixel transfer connection before said pixels are loaded into the neighborhood extraction means for processing.
Claim 10
Dependent Claim
150 words
A parallel partitioned serial neighborhood processing system comprising: means for partitioning an image matrix into a plurality of contiguous segments; means for feeding contiguous segments to adjacent serial neighborhood transformation stages in a raster scan line format; said serial neighborhood transformation stages including neighborhood extraction means for sequentially accessing a central pixel and its surrounding pixels in the matrix, neighborhood transformation logic means for transforming the value of the central pixel as a function of its value and the values of its surrounding pixels contained in the neighborhood extraction means; a single connection for pixel transfer between adjacent stages; means for sensing the relative position of each pixel in the lines fed to each stage; and transfer control means for bi-directionally transferring pixels lying on the edges of each segment to adjacent stages over said pixel transfer connection before said pixels are loaded into the neighborhood extraction means for processing.
Claim 11
Dependent Claim
16 words
The system of claim 6 wherein said pixel transfer connection is an eight bit data bus.
Claim 12
Dependent Claim
146 words
The system of claim 6 wherein said transfer control means comprises: multiplexer means having a plurality of inputs, a first input being connected to said pixel transfer connection from a left-hand adjacent stage, a second input connected to the pixel transfer connection between a right-hand stage, and a third input being connected to a supply of pixels from the image segment fed to that stage; latch means connected to the output of said multiplexer means, the output of said latch being coupled to said neighborhood extraction means, to said left-hand pixel transfer connection through a first gate, and to said right-hand pixel transfer connection through a second gate; an address counter whose contents are a function of the relative position of each pixel in the lines fed to each stage; and wherein said address counter is utilized to control the multiplexer and first and second gates.
Claim 13
Dependent Claim
39 words
The system of claim 8 wherein said feeding means is operative to delay the feeding of each line of pixels to the outer stages by a suitable amount so as to prevent simultaneous transfer of pixels within a stage.
Claim 14
Dependent Claim
42 words
The system of claim 8 wherein each stage includes a memory connected between the output of the latch and the neighborhood extraction means, operative to store a given number of scan lines and sequentially load said neighborhood extraction means with neighboring pixels.
References Cited (Backward Citations)
US3106698A
Cited
US4174514A
Cited
US4215401A
Cited
US4224600A
Cited
US4395699A
Cited
US4395700A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1987-11-20 | 1988-05-21 | 1988-11-20 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1991-11-20 | 1992-05-21 | 1992-11-20 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 1995-11-20 | 1996-05-21 | 1996-11-20 | $7700.00 | Window Closed / Lapsed |