JPH08235332A - Dimensional characterization method of boundary line-type buried data block - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method by which a boundary line type buried data block having an internal area containing no buried data can be dimensionally characterized without impairing the visual homogeneity of the data block. SOLUTION: A buried data block 21 is dimensionally characterized in disjoined sections surrounding an internal area by providing a grid-like synchronization frame at the center of nearly even intervals along a first set of parallel paths and a second set of parallel paths in the direction perpendicular to the first set of paths from one end section to the other end section including from the inner edges to the outer edges of the block 21 and encoding at least one address code in buried data characters 25 on the first set of paths and at least one additional address code in buried characters 25 on the second set of paths.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION This invention relates to record formats for self-clocking glyph codes, and more particularly to reading and / or interpreting glyph codes without compromising visual homogeneity. It relates to a record format that includes a control emoji infrastructure that encodes information that facilitates.

[0002]

2. Description of the Related Art "Data glyphs" (data glyphs; pictograms that encode user data) are called "synchronous pictograms" (s
ync glyph; The record format for the self-clocking emoji code, which spatially correlates with the additional emoji that spatially synchronizes the emoji reading process,
Proposals have been made. To that end, a "pictogram pattern" is recorded by writing data pictograms and synchronous pictograms to a recording medium according to a predetermined spatial format rule. Furthermore, since the sync glyphs are spatially dispersed within this glyph pattern according to a preselected spatial distribution rule, the positioning of the sync glyphs must follow a predefined geometric sub-pattern.

In order to obtain a visually uniform pictogram pattern, synchronous pictograms and data pictograms cannot be visually distinguished. In fact, all emoji are generally binary "0"
About + 45 ° from the vertical line to code and “1”
And symbols from the same set of symbols, such as diagonal symbols with -45 ° tilt.

However, in accordance with the teachings of the above proposals, sync glyphs have a predetermined sync code sequence such that the logical ordering of the bits of the "sync code sequence" is ensured by the spatial ordering of the sync glyphs encoding them. Encode consecutive bits of. Therefore, in order to identify these sync glyphs, the decode values of the glyphs spatially distributed within the glyph pattern that closely match the known sync glyph sub-pattern are examined. This check is performed on successive sets of glyphs until a large number of glyphs with decode values that are substantially correlated with the known sync code sequence are found (in practice, this correlation process allows a small number of correlation errors). Desirable). The goal is to establish that there is a correlation across a large number of glyphs and to reasonably confirm that a particular glyph is a synchronous glyph. Then, by extending the inspection of emoji decode values according to known sync emoji subpatterns and rules that determine how to map the emoji decode values into memory,
Identify additional sync emoji.

Typically, a sync glyph sub-pattern consists of a linear array of one or more sync glyphs. A linear array of intersecting synchronized glyphs can be used to spatially synchronize the glyph read / decode process in two dimensions (eg, along the x and y axes of standard Cartesian coordinates). , Attractive. The grid-like synchronized glyph framework also provides a suitable spatial reference for spatially synchronizing the glyph reading process in two dimensions, as well as the presence of localized damage and / or distortion in the glyph pattern. In some cases, it may be convenient to provide multiple paths to navigate from any one sync glyph to any other sync glyph via the sync grid, allowing recovery of spatial synchronization. There are many.

A "glyph" was defined as a two-dimensional image symbol having at least two graphical states to encode a single bit logical state ("1" and "0"). It is an "embedded data character". An "embedded data block (abbreviated as EDB, hereinafter)" is a two-dimensional image symbol for storing and retrieving data. The EDB (or embedded data block) consists of embedded data pictograms, some of which are coded to define synchronization frames and other of which are coded to support user / application specific information. The synchronization frame (sometimes referred to as the “pictogram synchronization sub-pattern”) and user information are 2 of the EDB.
There are two major structural components. However, in order to transfer information about the structural or logical organization of the EDB from the EDB's composer to the reader / decoder, it is necessary to include additional components, including both implicit logical and explicit graphical structures. It will be appreciated that the configuration of the EDB can be extended. "Emoji pattern"
Is an example of EDB.

In accordance with the present invention, a border-embedded data block can be decomposed with an address code attached to the embedded data characters in the sync frame, thereby allowing the data block to be dimensionally characterized by those address codes. To

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 shows a rectangular self-clocking pictogram code pattern 21 printed on a suitable recording medium 22, such as a normal plain paper document. For example, the code comprises diagonally-shaped pictograms 25, as can be clearly seen by the enlarged portion 23 of the pictogram code pattern 21. The pictograms 25 are "0" and "1" as indicated by reference numeral 27.
In order to encode the above, the recording medium 22 is inclined about + 45 ° and −45 ° with respect to the vertical axis. Since each code value is represented by the presence of glyphs, no data is coded in the space between glyphs 25, or at transitions at their edges. As a result, the pictogram 25 is properly printed in the substantially uniform center, so that the pictogram code 21
Has a nearly homogeneous visual appearance. In fact, the scale factor for printing the pictographic code 21 is often so small that the individual pictograms 25 are essentially mixed with each other when viewed with the naked eye in a normal observation state. This is especially so that the machine-readable digital data is visually inconspicuous,
It is also an important advantage for applications that require it to be aesthetically pleasing to embed in an image, or for applications that benefit from it.

As a practical problem, the pictogram code pattern 2
One or more dimensions of the embedded data block, such as 1, are of near optimal size for machine-readable message length embedded within the pattern, ie, code patterns that better meet geometric layout requirements. Can be a variable that can be changed from time to time. However, most (if not all) of the decoding processes that can be used to recover the embedded data will result in the expected number of emojis 25 that the data block 21 contains and the basic lossout of those emojis. In order to determine, the knowledge of the dimensions of the data block 21 with sufficient accuracy is relied upon. For example, in the case of a simple rectangular data block, decoding generally requires the number of pictogram columns in the code pattern and information specifying the number of pictograms in each column. As will be seen later, the layout of the pictogram 25 could be partially specified by constructing the data block 21 with a code frame of a certain size, but even after that, the data block 21 contains The size of the data block 21 can be changed by increasing or decreasing the number of contained frames, by spatially moving one or more frames, or both.

In accordance with the present invention, an address code (eg, a maximum length pseudo noise (PN) sequence code).
Provides a spatial synchronization reference as well as a distributed label (from which one or more unknown dimensions of the code pattern 21 can be determined) to generate a reference glyph contained within the code pattern 21. (Reference glyp
coded by h). The term "address code" as used herein accurately determines the position (ie, relative address) within a sequence of any given value by reading a specified number (m) of values within the sequence. It is a unique sequence of values that can be. Here, m is a sequence dependent parameter that identifies the minimum number of values that must be read to determine the position of a given value in the sequence. A "pseudo noise sequence" (sometimes called a "PNS" or PN sequence) is an example of an address code that can be used for most applications. A PN sequence is a sequence of binary values generated by an n-element shift register that is preloaded with an explicit seed value and operated with a feedback tap at a specific register location. A "maximum length PN sequence" is unique over a length of 2 ⁿ -1.

In particular, the size along the unknown dimension of the data block pattern (measured by the glyph count) is a pair of spatially correlated reference glyphs (aligned parallel to each other and the glyph pattern unknown). It will be understood that each can be calculated from the relative address of each of the reference pictograms (which are located in a linear array of reference glyphs).
In fact, in order to simplify the calculation, a linear array of these reference pictograms is not necessary,
Advantageously aligned parallel to one unknown dimension. As will be appreciated, it would be entirely successful to provide a data block system in which the block dimensions uniquely determine all the remaining structural variables needed to recover the embedded data from the code pattern or data block 21. Is.

Specifically, in accordance with one embodiment of the present invention, one of two arrays of reference pictograms is used to encode successive values of an address code sequence propagating from a spatial reference point at one end of the span. In contrast, the other array of reference glyphs encodes successive values of the address code sequence propagating from the spatial reference point at the other end of the span. Thus, the relative address of any given glyph in one array establishes the distance (within the glyph) to one edge of the glyph pattern, while the relative address of the spatially corresponding glyph in the other array. The address establishes a distance to the other edge of the glyph pattern (also within the glyph). Instead, when the reference glyph encodes successive values of a single address code sequence that propagates rasterically along one unknown dimension of the glyph pattern 21 from one corner of the pattern to the diagonally opposite corner, the From the relative address of a pair of pictograms that are spatially correlated like this,
It will be appreciated that the unknown dimension of the code pattern 21 can be calculated.

As will be appreciated, the size of the local area of the glyph pattern 21 that must be inspected is
(A) large enough to make sure that the reference pictogram (from which the relative address information is obtained) is laterally spatially correlated to the unknown dimension of the code pattern, and (b) It need only be large enough to ensure reliable identification of the relative addresses of those pictograms. In general, the boundaries of any of the glyph blocks for an address sequence need not be within the local area to be examined. In fact, these fiducial emojis may be obscured by other unrelated images, blurred for portions of the captured image for decoding, and due to available examples of emoji patterns. You may even lose it.

This distributed dimension labeling (distribu
ted dimensional labeling) because the dimensional information is spatially dispersed throughout the normal synchronization / address code, and the synchronization / address code is often spatially dispersed throughout the emoji code pattern 21. Very robust in terms of redundancy and distribution. As will be appreciated, distributed dimension labeling is especially relevant if some of the emoji blocks are missing due to capture, or if the emoji blocks are contained within an extended pattern of additional emoji, or both. Very useful in the case of.

In FIG. 2, the counter-propagating address codes U and V are shown.
3 shows the distributed dimension labeling obtained by encoding alternating linearly arranged glyphs according to successive values of. As shown, this reference pictogram sub-pattern extends transversely to the unknown dimension of the pictogram pattern 31 and at the same spatial frequency as the pictogram C that encodes the user data embedded in the pictogram pattern 31 (ie, , With the same center-to-center spacing). However, since the pictograms encoding the back-propagating address codes U and V are interleaved (interleaved) with each other, the sum of the relative addresses of the spatially adjacent reference pictograms (in this example, the spatial correlation is spatial Too close (spat
(equal adjacency)) multiplied by a scaling factor of k = 2 to determine the unknown dimension of the pictogram pattern 31.

For example, FIG. 3 shows a local area 34 of a data block 31 (not including any boundaries of the data block). From this local area 34, the dimensional information associated with the local address can be ascertained. In general, the address code is a maximum length PN sequence consisting of N bit long unique subsequences (ie, N stages of maximum length shift register code, etc.). As is known, a sample size of about 2N consecutive bits is recommended to reliably decode such codes. However, if there is no "ambiguity" in the code in use, then decoding can be performed using sample sizes as small as N consecutive correct bits.

More specifically, as shown in FIG. 3, the unknown dimension D _X of the data block 31 is (1) a distance variable X ₁ from the address code U to the left end of the measured data block 31, and (2) It is determined by calculating the sum of the distance variables X ₂ from the address code V to the measured right edge of the block and scaling appropriately. If the orientation "ambiguity" cannot be reliably analyzed by other methods, the underlying shift register codes for address codes U and V are conveniently separated. Also, different address codes can be used to carry additional distributed labeling information, as discussed in more detail below.

FIG. 4 shows the back-propagating address codes U and V.
The glyph pattern 41 by sequentially encoding glyphs of a linear array of two reference glyphs that are parallel, spatially adjacent, spread in a pattern according to successive values of
Shows the variance dimension labeling given as. As with the distributed dimension labeling approach described above, these interlaced address codes dimensionally characterize the glyph patterns 41 in a direction parallel to their propagation direction.
In fact, this interlaced dimensional labeling approach reduces the linear range of the local region that must be examined to recover the dimensional label and has a scaling factor k = 1 (ie, backpropagation in code pattern 41). The center spacing of the reference pictograms encoding the address codes U and V is the same as the center spacing of the data pictograms C), but is functionally similar to the interleaved dimension labeling described above.

A variation of interlaced dimension labeling is shown in FIG. In particular, backpropagating parallel address codes U, V are encoded by reference pictograms in non-adjacent arrays. Besides, this embodiment has the same basic structure as the embodiment of FIG. Although the lateral extent of the local region that needs to be investigated to recover the distributed dimension labeling information is increasing, the overhead required to obtain the synchronization criterion at a sufficiently high density is reduced by a factor of two. There is.

To interlace the address codes U and V, with special care, the relative addresses used to interpret the dimension labeling information belong to spatially correlated pictograms. You need to make sure. This correlation is typically 2 because a correlation process is used to identify and latch on the address code.
There is no guarantee that it will occur in spatially correlated places within one array. Fortunately, however, this potential "ambiguity" can generally be resolved using the self-clocking nature of emoji codes. For example,
If the address codes U, V are of maximum length PN sequence, the PN sequence U is read from the code block and compared with the complete PN sequence U, the start position of the segment Uss of the read PN sequence of the complete PN sequence U. Is determined. Furthermore, the spatial position Su of the start position having a correlation is recorded. The above procedure is then repeated for the PN sequence V in the code block to obtain the correlated position Vss and the corresponding spatial position Sv. Then, the position value of Vss is adjusted to generate spatially correlated pictograms for the read segments of the PN sequences U and V, for example according to the difference to the two spatial positions, and the PN sequence V is It is moved to a new starting position V'ss for reading.
Here, V'ss is the original start position Vss + (Sv-Su).
be equivalent to. Therefore, the sum Svu of V'ss and Uss is V
Equal to the size of the code block in the -U direction. If the PN sequences V, U are placed at locations other than all corresponding consecutive pictogram locations, the sum Svu is multiplied by the corresponding scaling factor used to place the original pictogram.

In addition to the above, FIG. 6 illustrates a non-parallel set of back-propagating address codes for multidimensional distributed labeling of pictogram code blocks 61. For this purpose, in FIG. 6, alternating back-propagation extending in the horizontal direction (X direction), alternating back-propagating to the address codes U and V in FIG. 5, extending vertically (Y direction). , Non-adjacent, interlaced, parallel address codes S, T
Is added. As will be appreciated, the two non-parallel sets of address codes provide distributed dimension labeling of the code block 61 in the X and Y dimensions as well as the X / Y spatial addressing of individual pictograms within the code block 61. And provide address / synchronization crosslinks to each other. As a practical matter, the pictograms encoding the address codes U, V usually extend at right angles to the pictograms encoding the address codes S, T, but this is not necessary.

As shown in FIG. 6, the vertical address codes S and T are the elements U where the horizontal address codes U and V intersect.
_00, a pictogram encoding a U _15, V _00, V _15, along with several other pictograms B, and is interleaved in the same way with a 50% duty cycle. This interleaving of at least one of the non-parallel sets of address codes U, V and S, T is convenient to avoid code conflicts at the intersections. Most regular interleaving patterns can be used that avoid potential conflicts between codes. However, a pattern that allows additional information to be encoded with some interleaved pictograms, such as pictogram B in FIG. 6, is convenient. For example, the pictogram B can be used to encode data useful in correctly decoding / interpreting the data block 61.

FIG. 7 illustrates the use of expansively Folded or raster-scanned extended address sequences for distributed dimension labeling. This is an alternative to using backpropagating address codes. As shown,
The pictograms in parallel sync frame rows encode each segment of the extended length address code (which can be considered folded over the sync signal frame rows). Note that other code interleaving can also be used. The vertical data block dimension for a sync frame row is determined by the difference in the relative address of the spatially corresponding pictograms on adjacent parallel sync frame rows (ie, in the middle sync frame row whose relative address is L). 1 / L of the difference, if they arise from spatially corresponding pictograms in sync frame rows separated from each other).

This distributed dimension labeling approach requires longer unique address code sequences for data blocks containing two or more parallel sync frame rows transverse to any given axis. This means increasing the size of the area that must be examined to ensure capture of the relative address of the spatially corresponding reference pictograms.

8-11 show general implementations of the invention for a number of different embedded data block types, such as the simple EDB 71 (FIG. 8) and border EDB 81 (FIG. 9). As shown, the EDBs 71 and 81 are shown in FIG.
72, a regular polygon (rectangle in this example) composed of linked frame blocks. The reason the frame block 72 is described as "linked" is that adjacent frame blocks have a common (ie, shared) border pictogram, such as 74 in FIG. In this unique implementation, each frame block 72 is 16 × 1
It contains an array of 6 glyphs, with the glyphs distributed on a center of essentially uniform spacing according to a regular grid pattern (the glyphs are represented by individual cells of the grid, such as 73 in FIG. 8). ). Therefore, the border pictograms of the frame block 72 jointly form a regular grid-like framework for each of the EDBs 71 and 81.

To begin with, describing the general features of the grid frames given to the EDBs 71, 81, the encoding for pictograms along each reach of these grids is the address code for the reader / decoder. It will be appreciated that it is the periodic interleaving with additional information that characterizes the EDB in detail. In the illustrated embodiment, the additional information explicitly coded by the glyphs in the grid frame includes key codewords and flags. These flags indicate whether or not the special processing codeword is found in frame block 72. The key codewords and special processing codewords will now be described in some detail.
However, at this point it is worth noting that the flags for special processing codewords are encoded by the pictograms at the intersections of the corners of the orthogonal grid frames, thereby causing address code conflicts (ie, along the orthogonal axis of the grid). Conflicts between extended address codes) and key code word conflicts (ie,
The interleave function is selected so as to avoid a conflict between key codewords regarding diagonally adjacent frame blocks.

As will be appreciated, distributed dimensional labeling, distributed block type labeling, and distributed rotationa for EDBs 71, 81.
Several different address codes are used to label (l orientation). However, in general, these address codes all have the same basic logical structure so that when encoded with pictograms at a fixed duty cycle, they map spatially consistently to the grid pictograms. For example, in this unique implementation,
Given that frame block 72 is a 16x16 emoji dimension, each address code is coded with a grid glyph at a 2/3 duty cycle (ie, two for every three glyphs in a grid frame). A pictogram encodes the address code), a 9 element maximum length shift register code (ie, 9 bit wide PNS) is suitable. At 2/3 duty cycle, the 9-bit long subsequence of the encoded address code spans 14 pictograms. It should be noted that the bits (eg, the initial bits) at corresponding positions within the continuous subsequence of coded address codes move relative to each other by 16 pictograms. Therefore, the use of interleaving with a 2/3 duty cycle does not require any glyphs at any intersection of the grid, and it is a 9-bit wide PN sequence or P
It will be appreciated that the NS consecutive subsequences are mapped to respective frame blocks 72 of EDBs 71, 81. Thus, these "corner" glyphs can be reserved for encoding flag bits, as previously described. The above interleaving is performed by a key code word and an error correction code (abbreviated as ECC) for protecting the key code word.
There are four additional pictograms on each side of each frame block 72 for encoding words.

Next, E performed by the address code
The distributed labeling of the DBs 71 and 81 will be described. The glyphs along adjacent parallel reach of the lattice frame encode the back-propagating address code. In this particular embodiment, neither EDB 71 or 81 is allowed to have more than 640 glyphs along its width (x dimension) or its height (y dimension). Therefore, considering the 2/3 duty cycle used to interleave the ad-race code with the key codeword and the special processing codeword bits, the use of 9-bit wide PNS for these address codes is It is fully guaranteed that the sub-sequence will not be repeated even in.
Therefore, the address code attaches the distributed dimension label to the EDBs 71 and 81. These dimension labels specify the width and height of the simple EDB 71 with pictograms, as described above. The dimension labels attached to the EDB 81 are likewise width and height specifications, but because of the more complex geometric shapes, these labels specify different widths and heights for different parts of the EDB 81. . Next, the subject will be described in some detail. However, EDB71 and E
The same 2 / for interleaving the address code of DB81
Since a 3 duty cycle is used, EDB7 is used to scale from the relative address given by the backpropagating address code to the dimension of the EDB measured pictogram.
A magnification of k = 3/2 is used for 1 and EDB81.

Specifically, the address code encoded by the pictograms on the adjacent x-direction reach of the lattice frame is not only (1) backpropagated, but (2)
x sync code (ie, all EDs used to distinguish x-axis from y-axis, regardless of EDB type)
B address code) is interleaved with an EDB identification code (ie, a type-specific address code that distinguishes each EDB type from every other EDB type). As shown, the x sync code propagates from left to right, with the first value of the code sequence being the leftmost unshared glyph within the reach (ie, the leftmost glyph not shared with the reach extending in the y direction). Coded. On the other hand, the EDB identification code propagates from right to left, and the first value of this code sequence is encoded in the rightmost unshared pictogram in the reach.

Similarly, the address code encoded with the pictogram on the adjacent y-direction reach of the lattice frame is:
In addition to (1) backpropagating, (2) the y sync code (ie, the address code of all EDBs used to distinguish the y-axis from the x-axis) is type-specific EDB
Alternately interlace with the identification code. As shown, the y sync code propagates from top to bottom, and the first value of the code sequence is encoded with the top unshared pictogram in its reach. Therefore, the EDB identification code back-propagates from bottom to top, and the first value of this code sequence is encoded in the top unshared pictogram in the reach.

In general, the address code is a maximum length PNS sequence that can be constructed with n element shift registers having different seed values and different register tap locations. Here, for example, the address code is
For (a) x sync code, [4, 9] register tap location, (b) for y sync code, [3, 4,]
6, 9] register tap location, (c) simple EDB type identification code [1, 4, 8, 9] register tap location, and (c) boundary EDB type identification code [2, 3] , 5, 9] can be configured using a 9-element shift register with register tap locations of Since it is desirable that the code does not contain long strings of "0" (lower frequency artifacts tend to cause visual degradation of the emoji code pattern), the seed used to construct the code above. The value is [001010101].

As shown in FIG. 9, the rectangular interior region 82 of the borderline EDB 81 is excluded from the dimension specifications given by the distributed dimension labeling of the EDB 81. To this end, the propagation of the address code along the grid reach glyph interrupted by the interior region 82 causes the last unshared glyph (ie, ED) to reach before being blocked.
It ends at the inner edge of B81). The propagation of the address code is then resumed on the other side of the internal area 82, and the first value of the sequence is encoded in the first non-shared pictogram contained in this second area of the EDB 81. In short, the borderline EDB 81 is effectively decomposed into a set of 4 contiguous simple EDBs by the distributed dimension label attached to the EDB 81. Each of the simple EDBs is dimensionally characterized by a distributed dimension label of the same type used to dimensionally characterize the simple EDB 71.

To characterize the type of EDB, the ED
Examine the EDB identification code embedded in B to obtain a known E
Contrast with DB type identification code. Similarly, to rotationally disambiguate the EDB, the x and / or y sync codes are identified, and if the code sequences need to be read in normal order (first to last), the EDB type Examine one or both of the codes, E
Reorient DB reads.

In this implementation, as shown in FIG. 10, the ECC that protects the key codeword and its keyword code is used.
(Error Correction Code) Four pictogram positions are reserved on each side of each frame block 72 to encode the codeword. In another implementation, a different number of glyph positions to interleave glyphs encoding logically ordered data values that define one or more common features of frame block 72 within EDB 71 or 81. Can be saved. Note that the data encoded by these interleaved pictograms is justified to be protected by one or more separately calculated ECC codewords for the data embedded in the interleaved pictograms. It may or may not be important.

With the above in mind, there is no contradiction of EDB.
It can be seen that the key codewords on the shared boundaries between adjacent frame blocks 72 must be reordered to maintain layout.

The key code word data begins at (i) the intersection of the x-direction sync line and the y-direction sync line and (ii) x
It is placed in several reserve locations in the clockwise direction starting at the intersection of the direction EDB identification row and the y-direction EDB identification row.

Since the pictogram encoding the address code functions as a pointer for the pictogram encoding the key codeword, the pictogram encoding the key codeword for the frame block 72 is the pictogram encoding the address code. Interleaving at facilitates quick and reliable recovery of key codeword data from frame blocks. The key codeword data read order for a given frame block is determined by the x and y positioning of the given frame block in EDB 71 or EDB 81 with respect to the upper left corner of the EDB. This is a reliable indicator of the read order of a given frame block 72.

Frame block 72 and / or E
Although the various types of data that characterize the DB 71 or 81 could be encoded by a key codeword, in the illustrated embodiment, the key codeword is used in the recovery of the data embedded in the EDB. It is used to provide information to initiate the error correction process (by means not shown). For this purpose, the first 7 bits of the key code word are 0-127 depending on the error correction code (ECC) word in the EDB.
Error correction level given by the scale of (ie,
It is used to specify the number of errors that can be corrected by these ECCs. In the Reed-Soloman error correction method, this specification is appropriately determined by the traditional rule that the number of errors that can be corrected is equal to 1/2 of the number of parity bytes in the ECC. The remaining bits of the key codeword are EDB71 or 81.
It is used to signal whether the ECC in is 128 bytes long or 256 bytes long (these are the two ECC byte sizes supported in this implementation).

Next, FIG. 11 will be described. A predetermined number of pictograms within each frame block 72 may be used to encode information regarding application-specific processing instructions and / or user-selectable processing instructions.
As will be appreciated, a reader / decoder (not shown) is predefined to map different encodings of some or all of the glyphs for these special processing instructions to specific operations. The information that these special processing instructions can provide can take many different forms because there may be mappings. As shown, the top row glyph within each frame block 72 (ie, a total of 14 glyphs) can be used to code the special instructions when needed. However, preferably, these pictograms can be used for normal user data encoding in the absence of special processing instructions. Thus, the pictograms in the corners of frame block 72 encode a flag state that indicates whether frame block 72 contains special processing instructions.

Similarly, interleaving the pictogram encoding the special processing flag condition with the pictogram encoding the address code facilitates quick and reliable recovery of the flag condition. However, since the flag states are unordered values, it is appropriate to code them with pictograms in the corners of data block 72. This is because there is no risk of inconsistency in ordering due to adjacent frame blocks sharing a corner pictogram horizontally or diagonally. In other words, the special treatment flag state is just one example of a type of "unordered information" that can be coded by a corner glyph.

[Brief description of drawings]

FIG. 1 shows an embedded data pattern consisting of a pictogram code pattern encoding machine-readable binary information, its expanded fragment, and the corresponding interpretation.

FIG. 2 maps an interleaved linear backpropagating address code onto a synchronous pictogram in a pictogram code pattern to dimensionally characterize the code pattern along an axis parallel to the direction of code propagation. It is a figure which shows operation.

FIG. 3 illustrates dimensional labeling performed in the embodiment of FIG.

FIG. 4 is an interlaced linear backpropagation onto the synch pictograms in adjacent parallel rows of the pictogram code pattern to dimensionally characterize the code pattern along an axis parallel to the direction of code propagation. It is a figure which shows the operation which maps the address code.

FIG. 5 is a diagram similar to FIG. 4 except that backpropagating address codes are mapped onto synchronization pictograms in adjacent non-parallel rows of the pictogram code pattern.

FIG. 6 is a schematic view of a back-propagating address code for dimensionally characterizing a pictographic code pattern in a two-dimensional space.
FIG. 6 is a diagram illustrating the operation of mapping each backpropagating address code set onto a two-dimensional framework of synchronous pictograms within a pictogram code pattern.

FIG. 7: In order that the mapping of the address code characterizes the emoji code pattern in the two-dimensional space dimensionally, the raster of each address code is folded back on the two-dimensional framework of the synchronous emoji in the emoji code pattern. It is a figure which shows the operation which maps to a shape.

FIG. 8 is a partial view of a simple embedded data block consisting of a plurality of frame blocks in the format shown.

9 is a schematic diagram of a borderline embedded data block consisting of a plurality of frame blocks of the type shown in FIG. 8. FIG.

FIG. 10 is a schematic diagram showing permuted placements for key codewords and reordered reads, consisting of bit sequences shared by key codewords for adjacent frame blocks.

FIG. 11 is a schematic diagram of a frame block having a flag in its sync frame set that indicates that the frame block has an optional special processing codeword.

[Explanation of symbols]

 21 pictographic code pattern (data block) 22 recording medium 23 enlarged part of pictographic code pattern 21 25 marking with diagonal lines or "pictogram" 27 figure showing interpretation of enlarged part 23 pictographic pattern (data block) 34 local of data block 31 Area 41 Pictogram pattern 61 Pictogram code block (data block) 71 Embedded data block (EDB) 72 Linked frame block 73 Individual cell of lattice 74 Shared border pictogram 81 EDB 82 Internal area of border type EDB 81

Claims (3)

[Claims] 1. A method for dimensionally characterizing a border-type embedded data block in which there is no embedded data in an internal region, the edge-to-edge of the embedded data block including from an inner edge to an outer edge. Up to the part, along the first set of parallel paths and along the second set of parallel paths in a direction orthogonal to the first set of paths, at the center of the substantially uniform spacing of the embedded data characters. Providing a grid-like synchronization frame, encoding at least one address code in the embedded data characters on the first set of paths, and encoding at least one additional address code in the embedded data characters on the second set of paths. And enabling the data block to be dimensionally characterized in a decomposed portion surrounding the interior region. 2. The method of claim 1, wherein the embedded data block comprises a self-clocking pictogram code pattern. 3. The method according to claim 2, wherein the pictographic code pattern is composed of slanted symbols, and the slanted symbols are written on the recording medium at centers of substantially uniform intervals. The method of claim 2.