MXPA96000190A - Method for recognition of writing entry man - Google Patents

Abstract

The present invention relates to a method characterized in that it comprises the steps of: receiving the input of handwriting characters, that input of handwriting characters is comprised of at least a first and a second discrete continuous segments, calculating at least one distance between the discrete continuous segments, first and second, the distance being essentially parallel to the writing axis of the first and second discrete continuous segments, identifying which of the plurality of distances has the smallest value, and using the distance to determine whether the segments continuous discrete, first and second, belong to separate handwriting character entry

Description

'METHOD FOR RECOGNITION OF MANUAL WRITING INCOME FIELD OF THE INVENTION This invention relates in a general way to the handwriting recognition, and, more particularly to the recognition of individual words.

BACKGROUND OF THE INVENTION The recognition by handwriting machine of people is a very difficult problem, and with the recent explosion of computing devices based on a pen, it has become an important problem to solve. The recognition by handwriting machine of people has various current applications. An example of the normal application for handwriting recognition of people is found in personal digital assistants, such as the EO and Newton products. Normally, these types of products have a touch-sensitive screen on which a user can write manually. Thus, these devices work to digitize what has been written manually, such as alphanumeric data, and then be able to process that data entered, in order to try to recognize the information contained in the € _ "manual frying. According to a handwriting recognition technique of the prior art, one performs a better determination as the identity of each alphanumeric character in sequence, the resulting character string comprises the result of the recognition activity. There are a variety of disadvantages to this approach. It is hampered by the difficulty of identifying the spatial limits of candidate income (in this case, alphanumeric characters that will be recognized). When these limits are not located correctly, it is impossible to accurately recognize the character, since it will lack parts or incorporate foreign material from adjacent characters. A significant problem with machine recognition of manual handwriting is the ability to recognize the end of an entry and the start of the next entry. For example, there is a significant problem in locating the end of a handwriting entry segment, word entry or alphanumeric, from the beginning of the subsequent handwriting entry segment, word entry or alphanumeric entry. Poor recognition of these separations in the manual writing income results in a poor and imprecise interpretation of the computer content of the writing input Tfey ial. In accordance with the foregoing, there is a need for a manual handwriting recognition technique, which can detect the end or end of a first handwriting entry segment of the beginning of a second handwriting entry segment, in the entry of handwriting and therefore provide a more accurate interpretation of the computer content of the writing entry - "" annual.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates an operation flow diagram in accordance with a preferred embodiment of the present invention. Figure 2 illustrates a graphical view of an illustrative representation in accordance with a preferred embodiment of the present invention. Figure 3 illustrates a graphical view of an illustrative representation in accordance with a preferred embodiment in the present invention. Figure 4 illustrates a graphical view of an illustrative representation in accordance with an alternative preferred embodiment of the invention. Figure 5 illustrates a graphic view of an illustrative representation of conformance to a modality pferred in the present invention.

DETAILED DESCRIPTION OF THE PREFERRED MODALITIES Normally, the input of a handwriting character is collected from the user in the form of discrete continuous segments. A continuous discrete segment consists of one or more pen strokes, where a pen trace is the mark left by a pen during the contact period -with an input device such as a paper or digitizing tablet. A trace is represented as a sequence of points sampled at approximately regular intervals by the input device. Each point is described by at least one X coordinate and one Y coordinate. The traces can be captured electronically using a digitizing tablet, or in an alternative mode they can be derived from a scanned image or faxed through a process line detection in the image; These methods of capturing an income electronically are understood in the art. Generally speaking, the present invention as determined determines whether two discrete continuous segments form part of the same entry of the handwriting character or part of more than one entry of a handwriting character. In the present invention, one or more discrete continuous segments are the handwriting entry units that "• • • • • • • • • • • • • • • • • • • • • • • • • • • Handwritten entry is the income that is captured electronically and that includes but is not limited to the following: handwriting entry, electronic entry, income captured by means of pressure, such as sealed or stamped income, the income that is received electronically, such as by facsimile, message generator, or other device For example, the present invention determines whether two continuous segments -discretes are part of the same word or are part of the same word. Separate Words In a preferred method, the present invention calculates one or more substantially parallel distances located substantially parallel to the writing axis, and compares these distances with one or more predefined thresholds.The predefined thresholds specify the minimum distance measurements that must be exceeded. by the substantially parallel distances so that the discrete continuous segments are judged as pe They belong to a separate manual writing entry, for example, separate handwriting words. The write axis is the line along which the manual write input is added. The write address is the address in which each subsequent manual write entry is added. In English, the handwriting entry is usually added along a horizontal writing axis and each subsequent alphanumeric entry follows horizontally after < S »previous entry in a writing direction that goes from left to right. Various other writing axes and alternative writing directions are possible with the implementation of the teachings of the present invention. In a preferred embodiment, the manual writing axis is horizontal and the manual writing input forms a series of words. In this preferred embodiment, the substantially parallel distances are calculated horizontally, and the output says whether the discrete continuous segments belong to separate words. In an alternative preferred embodiment, the handwriting axis is horizontal and the handwriting entry forms a series of separate characters, which may be alphanumeric characters, ideographic characters as found in languages such as Chinese, or other character forms or symbols of written communications. In this alternative mode, the output says whether the discrete continuous segments belong to separate characters. In another preferred embodiment, the manual writing axis is vertical and the manual writing input forms a series of separate characters, which may be alphanumeric characters, ideographic characters, or other manual writing text. In this preferred embodiment, the writing axis is vertical and the substantially parallel distances are aligned vertically; the output says if the continuous segments < * "Vretos belong to separate characters." Still in another preferred embodiment, the manual writing axis is vertical and the manual writing input forms a series of separate words, alphanumeric entry, or other manual writing input, such as a vertical list of words, or numbers In this preferred embodiment, the writing axis is vertical and the substantially parallel distances are aligned vertically; the output says whether the discrete continuous segments belong to separate handwriting entries, such as separate words. As discussed above and as will be discussed below, the present invention demonstrates through the display of various preferred embodiments that the writing axis can exist at any angle, and the manual writing input can be interpreted more generally as corresponds to discrete elements (including, but not limited to characters and words) that contain one or more discrete continuous segments. The application of the methods described herein to any of several preferred embodiments only requires a change in the coordinate system used and these modifications can be made in accordance with the teachings presented. Referring now to Figure 1, a preferred method of the present invention is illustrated. The present invention is applicable to one or more writing receipts ia "ifeual of discrete continuous segments. The preferred embodiments of the present invention are applicable to two or more entries of discrete continuous segment manual writing. The use of only two discrete continuous segments S ^ and S2 is for illustrative purposes. In the preferred method illustrated in Figure 1, the handwriting entry consisting of two discrete continuous segments S ^ and S2 (110) is accepted by a device, such as a PDA, or other device. Other devices that work to receive handwriting input include, but are not limited to, the following: computers, modems, message generators, telephones, digital or interactive televisions or others, devices that have a digitizing tablet, facsimile devices, scanned or scanned, and any device with the ability to capture manual writing income. Preferably, with the acceptance of the handwriting entry, the substantially perpendicular limits b ^ and? > 2 (120) between the discrete continuous segments S ^ and S2 • The limits b ^ and b2 that are substantially perpendicular to the writing axis are determined by finding the point in the sequence of the trace S ^ that has the longest offset along the the write direction and the point in the sequence of the trace S2 which has the smallest displacement along the "Writing lesson (120). The substantially perpendicular limit bi is the displacement value of the point in the sequence of the trace S ^ which has the largest displacement along the writing direction. The substantially perpendicular limit b2 is the displacement value of the point in the sequence of the trace S2 that has the smallest displacement along the writing direction. By calculating the displacement in the writing direction of each point in S ^ in turn, and comparing that value with a stored value which is initially a very large negative number, one can determine the substantially perpendicular limit b ^ that has the largest displacement along the writing direction. If the calculated offset value is larger than the stored value, the stored value is replaced with the calculated offset value. Once all the points have been examined, the stored value will contain the value of the largest displacement value found in the trace sequence. A similar procedure, starting with a very large positive start value, can be used to discover the point in S2 that has the smallest displacement along the writing direction by calculating the offset in the writing direction of each point in S2 to turn, and compare that value with a stored value that is Especially a very large positive number. If the calculated offset value is less than the stored value, the stored value is replaced with the calculated offset value. The preferred method then calculates a first substantially parallel distance d ', where d' = b - bx (130). Referring to Figure 1, the first substantially parallel distance d 'is compared to a first threshold -determined t '. If d 'is greater than or equal to (> =) t1, the preferred method concludes that the first continuous discrete segment S ^ and the second continuous discrete segment S2 belong to different manual write entry segments (145). By way of example, if d1 is greater than (>) t 'the preferred method concludes that S ^ and S2 are words, characters, or other elements other than handwriting entry. The selection of a predetermined threshold t 'is a value made in accordance with the specific modality. The selection of a threshold t 'will be discussed in more detail later. If the value of d 'is smaller, or smaller than, the threshold value t', additional processing is presented. To determine whether the discrete continuous segments S ^ and S2 form discrete continuous segments separated from the handwriting entry, such as words, characters, or other separate elements, a second distance is calculated bx. "substantially parallel d." The second substantially parallel distance d "is found by calculating several substantially parallel distances and selecting the shortest of those distances. In a preferred embodiment, this is done by first calculating the maximum perpendicular extension with the writing axis suspended by S ^ and S2 together (150). The extension is then divided into a number of bands of equal height substantially parallel to the axis - < you write (160). Then the substantially parallel distance between S ^ and S2 is found for each band (170) The shortest or smallest substantially parallel distance between S ^ and S2 of the plurality of bands is selected as d "(180) In the preferred method shown in Figure 1, once the substantially horizontal distance d ", the distance d" and the first substantially horizontal distance d 'combine to help make the final decision that S ^ and S2 are part of the same handwriting entry or are parts of discrete continuous segments separated from the handwriting entry A weighted average of d 'and d "is calculated. A preferred weighted average equation is q'd '+ q "d" / (q' + q ") By empirical testing on real handwriting data it has been found that a weighted average of d 'and d" is more I need either of the two alone. In a preferred embodiment, q '= q ", but If a judicious selection of q 'and q ", which are not identical, it may be possible to optimize this additionally, additionally, q' and q" may each be equal to zero, but not at the same time. The weighted average is compared against a second threshold t. "If the weighted average is less than or smaller than t", the preferred method concludes that S ^ and S2 are part of the same handwriting entry segment (197). If the weighted average is greater than or equal to t ", the preferred method concludes that Si and S2 are continuous discrete segments different from the manual write input and the preferred method concludes with that result (199) .The preferred method of selecting t 'and t "is a detail of the specific modality. In one modality, t 'can be adjusted by measuring the value of d' through a large data set and selecting the value that best distinguishes the real dividing points from the false selections. In another preferred embodiment, t 'can be adjusted to a constant value corresponding to some fraction of the distance between the input guides in the input device. In another preferred embodiment t 'may vary dynamically according to a fraction of the measured extent of S ^ and S2 perpendicular to the writing axis. In another modality, t 'can be explicitly adjusted by the user before qt t invoke the method. In the preferred embodiment illustrated, t1 is set to be 1/3 of the measured extent of Si and S2 perpendicular to the writing axis, which provides a good result for English alphanumeric text. Many other ways to adjust t1 can be considered and may be optimal for different solutions. The various ways of determining t "in different modalities are similar to those described for determining." In Figure 2, a graphic representation is provided for a preferred embodiment In this embodiment, the discrete continuous segments, Si (210) and S2 (220) correspond to words in English, the writing axis (230) is horizontal, and the writing direction (270) goes from left to right In this preferred embodiment, the movement along the writing direction for any point it is simply the value of the X coordinate for that point, so that the substantially perpendicular boundary bi (240) is the value of the X coordinate of the furthest right point in Si (210), and the substantially perpendicular boundary b2 (250) is the value of the X coordinate of the furthest point in S2 (220) The substantially parallel distance d '(260) is the distance extending between b and b2 Alternatively, as illustrated in n Figure 3, the discrete continuous segments Si (310) and S2 j * 0), correspond to Chinese characters, the writing axis (330) is vertical, and the writing direction (370) goes from the top to the bottom. In this mode, the displacement along the writing direction (370) for each point is simply the value of the Y coordinate for that point, multiplied by -1 (assuming a standard coordinate system in which Y increases as one moves from the bottom to the top), so that the substantially perpendicular boundary bi (340) is the value of the Y coordinate for -1 of the lowest point in Si (310), and the substantially perpendicular boundary b2 (350 ) is the value of the Y coordinate by -1 of the highest point in S2 (320). In other modalities, the values for bi and b2 can be calculated by applying a simple geometric rotation of the manual write input to align the writing axis with one of the cardinal axes of the coordinate system, and then apply the procedure just described. This is a direct operation that will be understood by people along with the teachings presented here. Alternatively, other modalities can calculate the displacements of the points along the writing direction, when the writing direction does not align with one of the cardinal axes of the coordinate system, by means of a simple geometric projection of the discrete continuous segments in - writing axis, and then using trigonometry to calculate the displacement of the points projected from the origin of the coordinate system. This is a well-understood mathematical procedure that can be applied to the teachings of the present. As previously stated, if d 'is smaller than t', processing continues. In many cases, the segments that in fact belong to separate words or characters are placed in such a way that d 'is small or even negative, but it is still possible to determine that Si and S2 belong to different words or characters. Figure 4 shows an example of this situation for a modality in which Si (410) and S2 (420) are English words. In this example, d '(460) is really negative, and bi (440) is greater than 2 (450). In this example, the writing axis (430) is horizontal and the writing direction (470) goes from left to right. In a modality in which the writing axis is horizontal, the maximum extension perpendicular to the writing axis subtended by Si and S2 is calculated by finding the maximum and minimum values of the Y coordinate present in Si and S2. To find the minimum value of the Y coordinate, a stored value starts with a very large positive value, and then each point in Si and S2 is examined in turn, and if the value of the Y coordinate is smaller than the value to. The stored value is assigned to that value of the Y coordinate. A similar method is used to find the maximum value of the Y coordinate. By subtracting the maximum and minimum values from the Y coordinate, the perpendicular extension can be calculated with The axis of writing All the points in S and S2 fall within this extension In a preferred embodiment, this process can be made more efficient by considering only a subset of the points in Si and S2, so that only a number is considered. fixed points near the limit in question, because these are the ones that most likely affect the measurement being derived.In an alternative modality, in which the writing axis is vertical, the maximum extension perpendicular to the axis of writing suspended by Si and S2 is calculated by finding the minimum and maximum values of the X coordinate present in Si and S2.To find the minimum value of the X coordinate, a stored value is starts with a very large positive value, and then each point in Si and S2 is examined in turn, and if the value of the X coordinate is smaller than the stored value, that value of the X coordinate is assigned to the stored value. A similar method is used to find the maximum value of the X coordinate. By subtracting the maximum and minimum values from the X coordinate, the extension perpendicular to the axis of the X coordinate can be calculated.

Jritura. In an alternative embodiment, where the writing axis is neither horizontal nor vertical, the plane of the image can be rotated to align the writing axis with the X or Y axis of the coordinate system, as described above in the calculation of the substantially parallel distance d 'when the writing axis was neither horizontal nor vertical. Once the extension perpendicular to the writing axis has been found, a preferred embodiment divides this extension into bands substantially parallel to the writing axis, such that each band describes a narrow band through Si and S2. Within each band, you find the point in Si with the largest displacement in the writing direction, and you find the point in S2 with the smallest displacement in the writing direction. This can be effected efficiently with respect to S by starting the stored value for each band with a very large negative number. Each point in Si is verified in turn. First, it verifies its perpendicular displacement with the writing axis to see which band falls inside. Because the bands extend to the full extent of Si and S2 perpendicular to the writing axis, it is guaranteed that a band can be found for any point in Si and S2. If the preferred modality described Internally to calculate the perpendicular extension by considering only a subset of points in Si and S2, the perpendicular extension may not contain all the points in Si and in S2, and so each point must be verified to ensure that it falls within a band. If not, it is no longer considered in an additional way. Once the band is identified, the displacement of the point in the writing direction compares the value stored for that band, and if the displacement is greater, it is assigned to the stored value. Once all the points in Si have been verified in this way, the value stored in each band will contain the largest displacement of the point with the largest displacement found in that band in Si. If there were no points in Si that are within a given band, the stored value will remain as the very large negative start value. A similar procedure, which uses a second value stored in each band started with a large positive number, is used with respect to S2 to find the minimum displacement of the point with the minimum offset in the writing direction within each band at S2. If there are no points for a given band in S2, the second value stored for that band remains at its large initial value. At the end of this operation, for each band there are two stored values corresponding to the displacement tr- large in the writing direction of Si in the band and the smallest displacement in the writing direction of S2 in the band. If the stored value has not changed from its very large positive or negative initial value, then the information for that band is not used in any additional calculation. Otherwise, a distance for each band is calculated by subtracting the stored value for the displacement of the point in Si of the stored value for the displacement • of the point in S2. The smallest of these distances is found by examining the distance for each band in turn and storing the smallest one found. This smaller distance is the substantially horizontal distance d ". If d" is less than 0, which can occur if S and S2 are touched or overlapped, the value is assigned 0. A mode in which the axis of writing is horizontal and the direction of writing goes from left to right, it is shown in Figure 5. In this mode, the number of bands (510) is 10. It is the point to the right of Si that falls within each band ( 520), and it is also at the furthest point of S2 that falls within that band (530). This is done with respect to Si by starting a stored value for each band at a very large negative value. In turn, each point is verified in Si. First, check your Y coordinate to see which band you fall into. Because the bands extend the overall height, . extension (540) of S and S, it is guaranteed that a band can be found for any point in Si and S2. Once the band is identified, the X coordinate of the point is compared to the value stored for that band, and if the X coordinate is larger, it is assigned to the stored value. Once all the points in Si have been verified in this way, the value stored in each band will contain the X coordinate of the furthest point found in that -you in Si- If there were no points in Si that fell within a given band, the stored value will remain as the very large negative start value. A similar procedure with respect to S2, using a second value stored in each band started with a large positive number, is used to find the X coordinate of the furthest point in each band S2. As previously described, the value of d "is calculated from the locations stored for each band The smallest or shortest substantially parallel distance, between Si and S2 of the number of plurality of bands is selected as d". The distance d "is then combined with d 'by a weighted average and compared with t". With the comparison of t ", the method of the present invention concludes whether the discrete continuous segments Si and S2 belong to the same manual write input or to separate manual write entries.

It will be apparent to those skilled in the art that the disclosed invention can be modified in numerous ways and can assume many different modalities from the preferred forms particularly disclosed and described above. In accordance with the foregoing, it is intended that the appended claims cover all modifications of the invention that are within the true spirit and scope of the invention and its -s-quivalents.

Claims (10)

1. .. 'NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and, therefore, the content of the following CLAIMS is claimed as property; A method characterized in that it comprises the steps of: receiving the input of handwriting-manual characters, that input of handwriting characters is comprised of at least a first and a second discrete continuous segments; calculating at least a substantially parallel distance between the discrete, first and second continuous segments; use the substantially parallel distance to determine whether the discrete, first and second continuous segments belong to separate handwriting character entries.
2. The method according to claim 1, characterized in that it comprises at least one of the following: a) the step of calculating at least a substantially parallel distance includes the step of computing a plurality of substantially parallel distances between the discrete continuous segments, first Y ^ .. rundo y, where also selecting the step of calculating a plurality of substantially parallel distances also includes the step of identifying the substantially parallel distance that has the smallest value and, where the step of using the distance is also selected parallel to determine whether the first and second discrete continuous segments belong to separate manual write-in income, the step of using the substantially parallel distance that has the smallest value is included to determine whether the discrete continuous, first and second segments belong to income of handwriting separated; b) further includes the steps of: identifying a substantially perpendicular limit for each of the discrete, first and second continuous segments; calculating a distance between the substantially perpendicular limit for the discrete continuous segments, first and second and, where in selecting also the step of using the parallel distance to determine whether the discrete continuous segments, first and second belong to separate manual writing entries, also includes the step of using the distance between the limit substantially ^ "* parapendicular to the discrete, first and second continuous segments to determine whether the discrete, first and second continuous segments belong to separate handwriting entries
3. 3. A method characterized in that it comprises the steps of: receiving the manual write input. , where the handwriting entry is comprised of at least - a first and a second discrete continuous segments, identifying a substantially perpendicular limit for each of the discrete, first and second continuous segments; calculating a first distance between the substantially perpendicular limit for the discrete, first and second continuous segments; when the first distance does not exceed at least the first predetermined threshold, calculate at least a substantially parallel distance between the discrete, first and second continuous segments; use the first distance and the substantially parallel distance to determine whether the discrete continuous segments, first and second, belong to separate handwriting entries.
4. 4. The method according to claim 3, characterized in that the step of using the first distance At a substantially parallel distance to determine whether the first and second discrete continuous segments belong to separate handwriting entries, it includes the step of adding a first weighted distance and a substantially parallel weighted distance to obtain a result and dividing the result by two to obtain a final result, and compare the final result with a second predetermined threshold.
5. The method according to claim 4, characterized in that it comprises at least one of the following: a) the predetermined thresholds, first and second, are identical to each other; b) the predetermined thresholds, first and second, are different from each other; c) a weighting factor of one, as applied to both the first distance and the second substantially parallel distance, to obtain the first weighted distance and the substantially parallel weighted distance, respectively, such that the first distance equals the first weighted distance and the substantially parallel distance is equal to the substantially parallel weighted distance.
6. 6. The method according to claim 3, characterized in that the step of calculating a first distance "The substantially perpendicular limit for the first and second discrete continuous segments includes the steps of: when the substantially parallel distance is positive in value, use the distance substantially parallel as the distance substantially parallel, when the substantially parallel distance is of negative value , using a predetermined value as the substantially parallel distance y, where when selecting also the predetermined value is zero 7.
7. The method according to claim 3, characterized in that the step of calculating at least a substantially parallel distance includes the step of calculating a plurality of substantially parallel distances between the first and second discrete continuous segments 8.
8. The method according to claim 7, characterized in that the step of computing a plurality of substantially parallel distances further includes the step of identifying substantial distances. parallel entities that have a smaller value and; where when selecting the step of using the first distance and the substantially parallel distance to determine whether the discrete continuous segments, first xj. second, they pertain to separate handwriting income, including the step of using the substantially parallel distance having the smallest value and the first distance to determine whether the discrete continuous segments, first and second, belong to separate handwriting entries.
9. The method characterized in that it comprises the steps of: receiving the handwriting entry, wherein the handwriting entry is comprised of at least a first and a second discrete continuous segments that are located substantially parallel to a writing axis; identifying a limit that is perpendicular to the writing axis for each of the first and second discrete continuous segments; calculating a first distance between the limit for the first and second discrete continuous segments; when the first distance at least exceeds a first predetermined threshold, it is concluded that the first and second discrete continuous segments belong to separate manual write receipts; when the first distance does not exceed at least the first predetermined threshold, calculate at least a substantially perpendicular distance between the segments First and second discrete tunnels, the substantially perpendicular distance is substantially perpendicular to the writing axis; use the first distance and the substantially parallel distance to determine whether the first and second discrete continuous segments belong to separate handwriting entries.
10. 10. A method for determining that two discrete segment-segments belong to separate manual write receipts, characterized in that it comprises the steps of: receiving the manual write input, the manual write input is comprised of at least one first and one second continuous discrete segments; identify a substantially perpendicular limit for each of the discrete continuous segments, first and second, the substantially perpendicular limits are those limits that will be more adjacent to each other; calculating a separation distance between the substantially perpendicular limit for the discrete, first and second continuous segments; when the separation distance exceeds at least a first predetermined threshold, it is concluded that the discrete continuous segments, first and second, belong to separate handwriting receipts; when the separation distance does not exceed at least the first predetermined threshold, calculate a plurality of substantially parallel separation distances between the portions of the discrete, first and second continuous segments; use the smallest separation distance of the substantially parallel separation distances, to determine whether the discrete continuous segments, first and second, belong to separate handwriting entries.