EP0238254A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Data compression using block list transform
Assignee: AT&T Inc
•
Filing: 1987-03-11
•
Granted: —
•
Expires: 2007-03-11
Abstract
Abstract A transform approach to image coding where pixels are coded in the order prescribed by a predetermined 'List'. The code for each pixel is developed by computing a prediction for the pixel based on the known pixel values in the neighbourhood of the pixel and subtracting this prediction from the true value of the pixel. This results in a sequence of predominantly non-zero valued pixel prediction errors at the beginning of the encoding process (of each block, and predominantly zero value pixel prediction errors towards the end of the encoding process. Enhancements to this approach include adaptive encoding, where pixel sequences having zero prediction errors are not encoded, and run length encoding of the signals to be transmitted or stored for augmenting the encoding process.
Inventors
- 👤 Barry Geoffry Haskell
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-02 11:39
CPC Classifications
H; H04; H04L; H04L12/00; H04N; H04N19/00; H04N19/50; H04N19/593; H04N19/60; H04N19/90; H04N19/93
Claims Analysis
Total Claims: 18
Claim 1
Independent Claim
111 words
A method for encoding data divided into blocks containing a plurality of data points, CHARACTERIZED BY the steps of: selecting a data point to be encoded; obtaining a set of prediction data points corresponding to said data point to be encoded, with said set of prediction data points, for most data points to be encoded, forming a neighborhood that is substantially centered about the corresponding data point to be encoded; developing a predicted value for said selected data point in accordance with said set of prediction data points; and developing a quantized prediction error signal from a quantized arithmetic difference between said data point to be encoded and said predicted value.
Claim 2
Independent Claim
111 words
A method for encoding data divided into blocks containing a plurality of data points, CHARACTERIZED BY the steps of: selecting a data point to be encoded; obtaining a set of prediction data points corresponding to said data point to be encoded, with said set of prediction data points, for most data points to be encoded, forming a neighborhood that is substantially centered about the corresponding data point to be encoded; developing a predicted value for said selected data point in accordance with said set of prediction data points; and developing a quantized prediction error signal from a quantized arithmetic difference between said data point to be encoded and said predicted value.
Claim 3
Dependent Claim
35 words
The method of claim 1, further comprising a step of developing a coded value for each said data point to be encoded that is related to said predicted value combined with said quantized prediction error.
Claim 4
Dependent Claim
34 words
The method of claim 2 further comprising a step of initially storing the data points of a block to be encoded; and a step of replacing initially stored data points with said coded value.
Claim 5
Dependent Claim
26 words
The method of claim 3 wherein said step of developing a predicted value combines said stored coded values specified by said set of prediction data points.
Claim 6
Dependent Claim
26 words
The method of claim 1 further comprising the step of encoding said quantized prediction error signal in accordance with a preselected variable word length encoding method.
Claim 7
Dependent Claim
58 words
The method of claim 1 wherein said steps of selecting a data point to be encoded, obtaining a set of prediction data points, developing a predicted value, and developing quantized prediction error signal are iterated for each data point in each of said blocks, and the sequence of selected data points to be encoded forms an ordered list
Claim 8
Dependent Claim
41 words
The method of claim 6 wherein said step of selecting a data point to - be encoded from said list adaptively progresses through said list in order or skips over preselected data points based on specific picture characteristics of said data.
Claim 9
Dependent Claim
49 words
The method of claim 6 wherein said list is an ordered list and said step of selecting a data point from said list selects data points in order when said quantized prediction errors are large, and selects data points out of order when said quantized prediction errors are small.
Claim 10
Dependent Claim
17 words
The method of claim 1 wherein most of said blocks include a data overlap with other blocks.
Claim 11
Dependent Claim
31 words
1 0. The method of claim 9 wherein said blocks of data represent image data and said data overlap appears along one horizontal edge and one vertical edge of said blocks.
Claim 12
Dependent Claim
156 words
A block list encoder for encoding datas CHARACTERIZED AS comprising: a data point memory (130) for storing input data point signals belonging to data blocks; a block list memory (140) for storing a list of data points in a block and storing for each data point a corresponding set of prediction data point designations; an address counter (160); means (157) for addressing said data point memory with said address counter and said block list memory; a predictor (120) responsive to output signals of said data point memory computation means (171, 173) responsive to output signals of said data point memory and to said predictor output signals for developing quantized prediction error signals; combining means (174) responsive to said computation means and to said predictor for developing coded values; means for storing said coded values in said data point memory (130); and output means (180) responsive to said computation means for delivering said quantized prediction error signals.
Claim 13
Dependent Claim
156 words
A block list encoder for encoding datas CHARACTERIZED AS comprising: a data point memory (130) for storing input data point signals belonging to data blocks; a block list memory (140) for storing a list of data points in a block and storing for each data point a corresponding set of prediction data point designations; an address counter (160); means (157) for addressing said data point memory with said address counter and said block list memory; a predictor (120) responsive to output signals of said data point memory computation means (171, 173) responsive to output signals of said data point memory and to said predictor output signals for developing quantized prediction error signals; combining means (174) responsive to said computation means and to said predictor for developing coded values; means for storing said coded values in said data point memory (130); and output means (180) responsive to said computation means for delivering said quantized prediction error signals.
Claim 14
Dependent Claim
19 words
The encoder of claim 11, wherein said output means comprise an encoder for compressing said quantized prediction error signals.
Claim 15
Dependent Claim
27 words
The encoder of claim 11, wherein said output means comprises a variable word length encoder and zero run length encoder for compressing said quantized prediction error signals.
Claim 16
Dependent Claim
11 words
The encoder of claim 11, wherein said predictor comprises averaging means.
Claim 17
Dependent Claim
28 words
The encoder of claim 11, further comprising means for inhibiting said output means; and means, responsive to said quantized prediction error signals, for controlling said means for inhibiting.
Claim 18
Dependent Claim
34 words
The encoder of claim 11, further comprising means for blocking said delivering of said quantized prediction error signals for sequences of said data points, based on preselected thresholds of said quantized prediction error signals.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.