FIELD OF THE INVENTION
[0001] The invention relates to a system of visual markers for fastening and digitizing
garment patterns and a method of digitizing garment patterns by using said visual
markers. More particularly, the invention relates to a digitization surface magnetized
by magnetic fasteners for arranging patterns and visual markers for identifying megadata
on said patterns during digitization. The visual markers can provide information on
the direction of a thread, snips or internal marks, a seam, auxiliary lines, building
lines, marks for relationship in plaid and striped fabrics, indicators of segments
of junction (stitching) between the patterns, among others.
BACKGROUND OF THE INVENTION
[0002] Traditionally, patterns used in the garment manufacturing industry were digitized
in a slow manner on a digitizing table where an operator identified various points
on the outline of the paper pattern by means of a crosshead and the system converted
said points into a polygon or a closed smooth curve by using such techniques as Beziers
and Splines.
[0003] With the advent of scanners, it became possible to automatically digitize paper patterns
as disclosed in
US7031527 that describes the use of large-size scanners, which is an expensive technology and
requires a large horizontal physical space.
[0004] The popularization of digital cameras allowed these high-performance devices to be
used in the industry and this invention aims at making use of said technology for
digitization of paper patterns. The challenge now is how to do it in a way that will
save physical space used for digitization. The use of walls to which patterns are
vertically fastened is a viable solution but makes the design more difficult.
[0005] The Certificate of Addition C1 0405039 discloses a magnetic pattern holder which
comprises a flexible supporting board that acts as a surface for supporting patterns
fastened through magnetic attraction by magnetized blocks.
[0006] The information on patterns is digitized by means of indications on the physical
pattern such as, for example, the direction of a thread, snips or internal marks,
stitching, auxiliary lines, building lines, marks for relationship in plaid and striped
fabrics, indicators of segments of junction (a seam) between the patterns, among others.
All of this information has a physical representation during pattern making and can
be automatically obtained.
[0007] However, patterns may have subjective characteristics, that is, information that
does not have a physical standardized representation on the pattern such as, for example,
snips that may have different representations in accordance with the pattern maker.
This specific and subjective characteristic makes the process of automatic digitization
of pattern more difficult.
[0008] In order to overcome these difficulties, this invention proposes visual marking optionally
associated with magnetized blocks, which allows additional information to be represented
in the pattern digitization.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009]
Figure 1 shows a circular magnetized block with visual marking, featuring the rotation
mark at its periphery.
Figure 2 shows 16 markers in a multi-bit system where the number of the supported
bits is 4 and is preferably used in this invention.
Figure 3 shows visual marking that identifies the rotation mark at the periphery and
Figure 3A shows a visual pattern determined by the internal visual mark as shown in
Figure 2 which is preferably used in this invention.
Figure 4 shows a representation of two visual markers to indicate the direction of
a fabric thread with relation to the pattern.
Figure 5 shows a flowchart of the steps to identify, classify, and correlate visual
markers with the patterns to be digitized.
Figure 6 shows a flowchart to determine the position of the visual markers in the
input image.
Figure 7 shows a flowchart of execution of the process that classifies the visual
markers in accordance with their type.
Figure 8 shows a flowchart of execution of the process that correlates the visual
mark of a certain type in the flowchart of Figure 6 with the digitized pattern.
DETAILED DESCRIPTION OF THE INVENTION
[0010] The system of visual markers for fastening and digitizing garment patterns, subject
matter of this invention, comprises a flexible board, magnetized blocks (20) for fastening
the patterns to the magnetic surface of the board and visual markers (30) for identifying
the megadata on said patterns (40) during digitization by means of a specific computer
software.
[0011] For a visual annotation of the patterns (40) placed on the surface of the magnetic
board such as, for example, the characteristics relative to the pattern making which
may not have a well-defined physical representation such as the direction of a thread,
snips, internal marks, stitching, auxiliary lines, building lines, among others, contrary
to the conventional manual textual descriptions which hamper understanding, this invention
describes visual markers (30) that provide a standardized manner of making the digitization
of this kind of annotations possible.
[0012] The visual markers (30) preferably have the property of being quasi-isotropic to
make the independent identification of the rotation easier.
[0013] In this invention, the visual markers (30) are circular with a single mark (31) at
the periphery as shown in Figure 1, said mark (31) allowing identification of the
, angle of rotation of the annotation with relation to any Cartesian plane.
[0014] By way of example within the scope of this invention, in order to make possible the
use of different visual markers (30) per set to be digitized, each visual marker (30)
has a visual standard belonging to a multi-bit alphabet (32) as shown in Figure 2.
The visual standard that identifies the type of annotation is internally printed on
the circular annotation and is capable of storing sixteen different types as shown
in Figure 2.
[0015] The visual marker (30) can be printed on one face of the magnetized block (20), on
two faces of the magnetized block (20) in order to maximize the number of usage options
or form a body disassociated from said magnetized block (20).
[0016] The following steps are carried out in order to digitize the visual markers (30)
arranged on the pattern (40) fastened to the surface of the magnetic board (10):
[0017] After inputting an image, the system identifies the exact position (100) of the visual
markers (30) in the image. The first piece of information necessary for digitization
is the exact position of the marker (30) in the input image. A cross-correlation algorithm
(120) that provides possible positions of the desired visual markers (30) is used
for this, said algorithm correlating an ideal computed image (110) of the visual marker
(30) to be identified with the input image of the system.
[0018] The cross-correlation algorithm (120) determines the exact positions of the center
of the visual markers (30) and a score associated with said position (130). In order
to increase the precision, this initial position is adjusted to possible neighboring
sub-pixels in the input image by using a bilinear interpolation algorithm.
[0019] Having the scores, a cutting algorithm is applied in order to remove possible positions
that did not score enough (140). These possible false-positive positions are found
by the cross-correlation algorithm (120).
[0020] With the central position of the precisely defined visual marker (30) and with the
information on the radius of the visual marker (30) to be detected, the single mark
(31) that determines the rotation (150) is found. The single mark (31) is used for
determining the angle of rotation of the visual marker (30) with relation to the horizontal
axis of the input image. Said single mark (31) also shows the point of interest of
the visual marker (160).
[0021] Having the information on the position and rotation of the visual marker (30), the
next step is to identify its type on the basis of the predetermined visual alphabet.
In the preferred embodiment of this invention, a four-bit visual standard (32) represents
said alphabet as shown in Figure 2. The classification step (200) determines this
value.
[0022] With the precisely determined center (130) and rotation angle (150) and having the
radius of the visual marker (30) that composes the binary code, the mean value of
the pixels (210) which composes each quadrant of the visual standard (32) is calculated.
[0023] Knowing the value of each quadrant, a binary value of size four (220) which is the
type associated with the visual standard (32) is calculated.
[0024] After determining the position and type of the visual markers (30), they are associated
with the digitized patterns (310) and stored together with the information on the
type of the visual markers (30).
1. A system of visual markers for fastening and digitizing garment patterns comprising
a flexible board, magnetized blocks (20) for fastening the patterns (40) to the magnetic
surface of the board and visual markers (30) for identifying megadata on said patterns
(4) during digitization by means of a specific computer software.
2. A system of visual markers for fastening and digitizing garment patterns of claim
1 wherein the visual marker (30) preferably has the property of being quasi-isotropic
with a single mark (31) at the periphery for identification of the angle of rotation
of the annotation with relation to any Cartesian plane and a visual standard (32)
that identifies the type of annotation.
3. A system of visual markers for fastening and digitizing garment patterns of claim
1 wherein the visual marker (30) is printed on one face of the magnetized block (20).
4. A system of visual markers for fastening and digitizing garment patterns of claim
1 wherein the visual marker (30) is printed on two faces of the magnetized block (20).
5. A method for digitizing garment patterns by using visual markers comprising the following
steps:
after inputting the image, the system identifies the exact position (100) of the visual
markers (30) in the image by using a cross-correlation algorithm (120) that provides
possible positions of the desired visual markers (30), said algorithm correlating
an ideal computed image (110) of the visual marker (30) to be identified with the
input image of the system;
the cross-correlation algorithm (120) determines the exact positions of the center
of the visual markers (30) and a score associated with said position (130);
a cutting algorithm to remove possible positions that did not score enough (140) is
applied;
with the central position of the precisely defined visual marker (30) and with the
information on the radius of the visual marker (30) to be detected, the single mark
(31) that determines the rotation (150) of said visual marker (30) with relation to
the horizontal axis of the input image is found;
identification of the visual standard (32) of the visual marker on the basis of a
predetermined visual alphabet in the classification step (200);
after determining the position and the type of the visual markers (30), said visual
markers (30) are associated with the digitized patterns (310) and stored together
with the information (320) on the type of the visual markers (30).