EP0926896A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Method for encoding wavelet coefficients
Assignee: Ricoh Co Ltd
•
Filing: 1998-12-22
•
Granted: —
•
Expires: 2018-12-22
Abstract
Abstract A method for encoding wavelet coefficients can achieve a high compression ratio for image data. The image data is converted and quantized into coefficients by a two-dimensional wavelet transform and the coefficients having the same spatial position are collected to form a block. A layered quad-tree structure of the coefficients included in each block is defined with respect to each of a set of coefficients representing a vertical component (LH), a set of coefficients representing a horizontal component (HL) and a set of coefficients representing a diagonal component (HH). Each quad-tree structure has a root corresponding to a coefficient of a lowest frequency band. A zero tree of which all descendant nodes lack a significant coefficient is searched for, and positional information of the zero tree is encoded. The coefficients other than descendants of the zero-tree are scanned so as to form a one-dimensional coefficient string of the scanned coefficients. The one-dimensional coefficient string is encoded according to a two-dimensional encoding method using a run length of a zero run immediately preceding a significant coefficient and a significant coefficient.
Inventors
- 👤 Masaaki Ishikawa
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 11:18
CPC Classifications
H; H04; H04N; H04N19/00; H04N19/60; H04N19/63; H04N19/64; H04N19/645; H04N19/647; H04N19/90; H04N19/96; H04N19/10; H04N19/102; H04N19/13
Claims Analysis
Total Claims: 10
Claim 1
Independent Claim
188 words
A method for encoding wavelet transform
coefficients in an image encoding method in which image
data is converted and quantized into coefficients by a
two-dimensional wavelet transform and the coefficients
having the same spatial position are collected to form a
block so as to encode the coefficients on an individual
block basis, characterized by the steps of: defining a layered quad-tree structure of the
coefficients included in each block with respect to each
of a set of coefficients representing a vertical
component (LH), a set of coefficients representing a
horizontal component (HL) and a set of coefficients
representing a diagonal component (HH), each quad-tree
structure having a root corresponding to a coefficient
of a lowest frequency band; searching for a zero tree of which all
descendant nodes lack a significant coefficient; encoding positional information of the zero
tree; scanning the coefficients other than
descendants of the zero-tree so as to form a one-dimensional
coefficient string of the scanned
coefficients; and encoding the one-dimensional coefficient
string according to a two-dimensional encoding method
using a run length of a zero run immediately preceding a
significant coefficient and a significant coefficient.
Claim 2
Dependent Claim
96 words
The method as claimed in claim 1,
characterized by the steps of: producing a zero-tree map having a layered
quad-tree structure corresponding to the coefficient
tree, each node of the zero-tree map indicating whether
or not a corresponding coefficient is a root of the zero
tree, each node being represented by one-bit data; scanning the zero-tree map according to a
depth-first method so as to change the zero-tree map
into a one-dimensional string of the nodes; and producing encoded data of the positional
information of the zero-tree map by encoding the one-dimensional
string of the nodes.
Claim 3
Dependent Claim
26 words
The method as claimed in claim 2,
characterized in that the one-dimensional string
produced from the zero-tree map is encoded by a run
length encoding method.
Claim 4
Dependent Claim
59 words
The method as claimed in claim 1,
characterized in that in the scanning step the
coefficients other than descendants of the zero tree are
scanned separately in each of the set of coefficients
representing the vertical component, the set of
coefficients representing the horizontal component, and
the set of coefficients representing the diagonal
component according to the depth-first method.
Claim 5
Dependent Claim
80 words
The method as claimed in claim 4,
characterized in that the set of coefficients
representing the vertical component is scanned in an
order of upper left, lower left, upper right and lower
right; the set of coefficients representing the
horizontal component is scanned in an order of upper
left, upper right, lower left and lower right; and the
set of coefficients representing the diagonal component
is scanned in an order of upper left, lower right, upper
right and lower left.
Claim 6
Independent Claim
233 words
A method for encoding wavelet coefficients
in an image encoding method in which image data is
converted and quantized into coefficients by a two-dimensional
wavelet transform and the coefficients
having the same spatial position are collected to form a
block so as to encode the coefficients on an individual
block basis, characterized by the steps of: defining a layered quad-tree structure of
coefficients included in the block with respect to each
of a set of coefficients representing a vertical
component, a set of coefficients representing a
horizontal component and a set of coefficients
representing a diagonal component, each quartered-tree
structure having a root corresponding to a coefficient
of a lowest frequency band; searching for a zero tree of which all
descendant nodes lack a significant coefficient; scanning the coefficients other than
descendants of the zero tree according to the depth-first
method, a scanning being separately performed for
each of the set of coefficients representing the
vertical component, the set of coefficients representing
the horizontal component and the set of coefficients
representing the diagonal component; producing a coefficient vector by using four
coefficients of each node of each quartered-tree
structure and producing an EOS code at a root of the
zero tree so as to produce a string of the coefficient
vectors and the EOS codes; and encoding the string of the coefficient vectors
and the EOS codes by a variable length encoding method.
Claim 7
Dependent Claim
80 words
The method as claimed in claim 6,
characterized in that the set of coefficients
representing the vertical component is scanned in an
order of upper left, lower left, upper right and lower
right; the set of coefficients representing the
horizontal component is scanned in an order of upper
left, upper right, lower left and lower right; and the
set of coefficients representing the diagonal component
is scanned in an order of upper left, lower right, upper
right and lower left.
Claim 8
Dependent Claim
63 words
The method as claimed in claim 6 or 7,
characterized in that in the encoding step the string of
the coefficient vectors and the EOS codes is adaptively
encoded by a variable length encoding based on a class
of a degree of a value of a coefficient corresponding to
a parent coefficient of the coefficient vectors or the
EOS code to be encoded.
Claim 9
Dependent Claim
55 words
The method as claimed in claim 6 or 7,
characterized in that in the encoding step the string of
the coefficient vectors and the EOS codes is adaptively
encoded by a variable length encoding based on a level
of a frequency band in which the coefficient or the EOS
code to be encoded is located.
Claim 10
Dependent Claim
63 words
The method as claimed in claim 6 or 7,
characterized in that in the encoding step the string of
the coefficient vectors and the EOS codes is adaptively
encoded by a variable length encoding using a
combination of a coefficient vector including a
significant coefficient or an EOS code and a number of
the coefficient vectors each of which lack a significant
coefficient.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.