 |
Claims  |
|
|
We claim:
1. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video image of at least one hairstyle;
processing the captured video image of the hairstyle to render it adaptable
to a variety of face sizes and shapes;
capturing the video image of a customer's face; and
combining the video image of the hairstyle with the video image of the
customer's face, in such a manner that the hairstyle image fits naturally
about the face image;
wherein the step of processing the hairstyle image is performed only once
for each captured hairstyle, and includes adding "fill" hair to the
hairstyle video image in a region covering a range of size and shape
adjustability; and wherein
the step of combining the hairstyle video image and the customer's face
image includes automatically deleting "fill" hair that would otherwise
cover the customer's face image, whereby the hairstyle video image is
combined with the customer's face image without unnatural gaps between the
customer's face image and the hairstyle image, without unnaturally
overlaying the hairstyle image over the customer'face image, and without
the need of operator intervention.
2. A method as defined in claim 1, and further comprising the steps of:
moving the hairstyle image with respect to the face image; and
repeating the combining step to be sure that the image have combined
without unnatural gaps or overlaying of hair on the face.
3. A method as defined in claim 1, and further comprising the steps of:
changing the color of the hairstyle image to a selected different color;
and
repeating the combining step to be sure that the images have combined
without unnatural gaps or overlaying of hair on the face.
4. A method as defined in claim 1, and further comprising the step of:
blending the colors of the hair and face images in a region close to a line
where the hair and face images meet.
5. A method as defined in claim 4, wherein:
the blending step is performed immediately after the combining step.
6. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video images of a plurality of hairstyles;
processing the captured video images of the hairstyles to render them
adaptable to a variety of face sizes and shapes, and to include a variety
of color versions;
capturing the video image of a customer's face;
selecting one of the hairstyles of which the video images have been
captured and processed; and
combining the video image of the selected hairstyle with the video image of
the customer's face, in such a manner that the hairstyle image fits
naturally about the face image;
wherein the step of processing the hairstyle images is performed only once
for each captured hairstyle, and includes adding "fill" hair to each said
selected hairstyle image in a region covering a region of size and shape
adjustability;
and wherein the step of combining the selected hairstyle image and the face
image includes automatically deleting "fill" hair that would otherwise
cover the face image, whereby the hairstyle image is combined with the
face image without unnatural overlaying the hairstyle over the face, and
without the need or operator intervention.
7. A method as defined in claim 6, and further comprising the step of:
blending the colors of the hair and face images in a region close to a line
where the hair and face images meet.
8. A method as defined in claim 7, wherein:
the blending step is performed immediately after the combining step.
9. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video image of at least one hairstyle;
processing the captured video image of the hairstyle to render it adaptable
to a variety of face sizes and shapes;
capturing the video image of a customer's face; and
combining the video image of the hairstyle with the video image of the
customer's face, in such a manner that the hairstyle image fits naturally
about the face image;
wherein the step of processing the hairstyle image includes adding "fill"
hair to the hairstyle video image in a region covering a range of size and
shape adjustability;
and wherein the step of combining the hairstyle video image and the
customer's face image includes deleting "fill" hair that would otherwise
cover the customer's face image, whereby the hairstyle video image is
combined with the customer's face image without unnatural gaps between the
customer's face image and the hairstyle image and without unnaturally
overlaying the hairstyle image over the customer's face image;
and wherein the step of combining the images further includes the following
steps performed for corresponding picture elements (pixels) in the two
images:
determining whether the hairstyle image pixel is either real hair, "fill"
hair, or neither,
determining whether the face image pixel is within the face area,
using the hairstyle pixel in the combined image is the hairstyle pixel is
of real hair,
using the hairstyle pixel in the combined image if the hairstyle pixel is
of "fill" hair and the face pixel is not within the face area, and
using the face pixel if the hairstyle pixel is of "fill" hair and the face
pixel is within the face area, or if the hairstyle pixel is of neither
real nor "fill" hair.
10. A method as defined in claim 9 and further comprising and the steps of:
moving the hairstyle image with respect to the face image; and
repeating the combining step to be sure that the images have combined
without unnatural gaps or overlaying of hair on the face.
11. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video image of at least one hairstyle;
processing the captured video image of the hairstyle to render it adaptable
to a variety of face sizes and shapes;
capturing the video image of a customer's face;
combining the video image of the hairstyle with the video image of the
customer's face, in such a manner that the hairstyle image fits naturally
about the face image; and
blending the colors of the hair and face images in a region close to a line
where the hair and face images meet, wherein the step of blending includes
selecting a face pixel immediately adjacent to a hair pixel,
for each primary color, determining the average primary color intensity for
the selected pixel and for all pixels adjacent to the selected pixel,
provided at least one of the adjacent pixels is a hair pixel,
replacing the primary color intensities indicative of the color of the
selected pixel with the averaged values of primary color intensities, and
repeating these steps of to lend the color of all face pixels adjacent to a
hair pixel;
wherein the step of processing the hairstyle image includes adding "fill"
hair to the hairstyle video image in a region covering a range of size and
shape adjustability;
and wherein the step of combining the hairstyle video image and the
customer's face image includes deleting "fill" hair that would otherwise
cover the customer's face image, whereby the hairstyle video image is
combined with the customer's face image without unnatural gaps between the
customer's face image and the hairstyle image and without unnaturally
overlaying the hairstyle image over the customer's face image.
12. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video images of a plurality of hairstyles;
processing the captured video images of the hairstyles to render them
adaptable to a variety of face sizes and shapes, and to include a variety
of color versions;
capturing the video image of a customer's face;
selecting one of the hairstyles of which the video images have been
captured and processed; and
combining the video image of the selected hairstyle with the video image of
the customer's face, in such a manner that the hairstyle image fits
naturally about the face image;
wherein the step of processing the hairstyle images includes adding "fill"
hair to each said selected hairstyle image in a region covering a range of
size and shape adjustability;
and wherein the step of combining the selected hairstyle image and the face
image includes deleting "fill" hair that would otherwise cover the face
image, whereby the hairstyle image is combined with the face image without
unnatural gaps between the face and the hairstyle and without unnaturally
overlaying the hairstyle over the face;
and wherein the step of combining the images further includes the following
steps performed for corresponding picture elements (pixels) in the two
images:
determining whether the hairstyle image pixel is either real hair, "fill"
hair, or neither,
determining whether the face image pixel is within the face area,
using the hairstyle pixel in the combined image if the hairstyle pixel is
of real hair,
using the hairstyle pixel in the combined image if the hairstyle pixel is
of "fill" hair and the face pixel is not within the face area, and
using the face pixel if the hairstyle pixel is of "fill" hair and the face
pixel is within the face area, or if the hairstyle pixel is of neither
real nor "fill" hair.
13. A method as defined in claim 12, and further comprising the steps of:
moving the hairstyle image with respect to the face image; and
repeating the combining step to be sure that the images have combined
without unnatural gaps or overlaying of hair on the face.
14. A method as defined in claim 12, and further comprising the steps of:
changing the color of the hairstyle image to a selected different color;
and
repeating the combining step to be sure that the images have combined
without unnatural gaps or overlaying of hair on the face.
15. A method as defined in claim 14, wherein the step of changing the color
of the hairstyle image including the steps of:
determining for the desired hair color a target average color intensity
level for each primary color;
determining for the selected hair color an average color intensity level
for each primary color;
determining for each primary color a multiplier value and an offset value,
which, when applied to the selected hair average color intensity level
will result in the target average color intensity level;
applying the multiplier and offset values to the primary color intensity
levels of each of the pixels making up the hairstyle image;
wherein the foregoing steps are performed for a plurality of selectable
hairstyles, as part of the step of processing the captured video images of
the hairstyles.
16. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video images of a plurality of hairstyles;
processing the captured video images of the hairstyles to render them
adaptable to a variety of face sizes and shapes, and to include a variety
of color versions;
capturing the video image of customer's face;
selecting one of the hairstyles of which the video images have been
captured and processed;
combining the video image of the selected hairstyle with the video image of
the customer's face, in such a manner that the hairstyle image fits
naturally about the face image; and
blending the colors of the hair and face images in a region close to a line
where the hair and face images meet, wherein the step of bending includes
selecting a face pixel immediately adjacent to a hair pixel,
for each primary color, determining the average primary color intensity for
the selected pixel and for all pixels adjacent to the selected pixel,
provided at least one of the adjacent pixels in a hair pixel,
replacing the primary color intensities indicative of the color of the
selected pixel with the averaged values of primary color intensities, and
repeating these steps to blend the color of all face pixels adjacent to a
hair pixel;
and wherein the step of processing the hairstyle image includes adding
"fill" hair to each said selected hairstyle image in a region covering a
range of size and shape adjustability;
and wherein the step of combining the selected hairstyle image and the face
image includes deleting "fill" hair that would otherwise cover the face
image, whereby the hairstyle image is combined with the face image without
unnatural gaps between the face and the hairstyle and without unnaturally
overlaying the hairstyle over the face.
17. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video image of at least one hairstyle;
predefining the hairstyle image to include "real" hair portions and "fill"
hair portions;
processing the captured video image of the hairstyle to render it adaptable
to a variety of face sizes and shapes;
capturing the video image of a customer's face;
predefining the video image of the customer's face to include a "face area"
portion that must not be dominated by "fill" hair elements of the
hairstyle image; and
combining the video image of the hairstyle with the video image of the
customer's face, in such a manner that the hairstyle image fits naturally
about the face image;
wherein the step of processing the hairstyle image includes adding "fill"
hair to the hairstyle video image in a region covering a range of size and
shape adjustability;
and wherein the step of combining the hairstyle video image and the
customer's face image includes deleting "fill" hair that would otherwise
cover the customer's face image, whereby the hairstyle video image is
combined with the customer's face image without unnatural gaps between the
customer's face image and the hairstyle image and without unnaturally
overlaying the hairstyle image over the customer's face image;
and wherein the step of combining the hairstyle image and the customer face
image further includes
displaying the "real" hair image in preference to corresponding elements of
the face image,
displaying the "face area" image in preference to corresponding elements of
the "fill" hair, and
displaying the "fill" hair image if not within corresponding elements of
the "face area".
18. A method for displaying hairstyle images on a video image of a
customer's face, comprising the steps of:
capturing the video images of a plurality of hairstyles;
predefining the hairstyle images to include "real" hair portions and "fill"
hair portions;
processing the captured video images of the hairstyles to render them
adaptable to a variety of face sizes and shapes, and to include a variety
of color versions;
capturing the video image of a customer's face;
predefining the video image of the customer's face to include a "face area"
portion that must not be dominated by "fill" hair elements of the
hairstyle image;
selecting one of the hairstyles of which the video images have been
captured and processed; and
combining the video image of the selected hairstyle with the video image of
the customer's face, in such a manner that the hairstyle image fits
naturally about the face image;
wherein the step of processing the hairstyle images includes adding "fill"
hair to each said selected hairstyle image in a region covering a range of
size and shape adjustability;
and wherein the step of combining the selected hairstyle image and the face
image includes deleting "fill" hair that would otherwise cover the face
image, whereby the hairstyle image is combined with the face image without
unnatural gaps between the face and the hairstyle and without unnaturally
overlaying the hairstyle over the face;
and wherein the step of combining the hairstyle image and the customer face
image further includes
displaying the "real" hair image in preference to corresponding elements of
the face image,
displaying the "face area" image in preference to corresponding elements of
the "fill" hair, and
displaying the "fill" hair image is not within corresponding elements of
the "face area." |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND OF THE INVENTION
This invention relates generally to graphics display systems and, more
particularly, to graphics display systems used to combine graphics images
from two or more source. The particular problem that the invention
addresses relates to the field of hair styling, but it will be apparent as
the description proceeds that the invention is equally applicable to other
fields, such as cosmetics, clothing fashions, and so forth.
A significant problem for hair stylists, and for their patrons or
customers, is the selection of a hairstyle that will look pleasing on a
particular person. Some hair stylists simply do not have the artistic
ability or the experience to anticipate the results of applying a
particular hairstyle to a customer. The customer will typically have even
greater difficulty in selecting a hairstyle from photographs and imagining
what the effect of the style will be. Because of differences in face
shape, facial features, type and color of hair, physical stature, and so
forth, a style that looks perfect on a photographed model may be totally
inappropriate for the customer.
In the past, the most common approach to solving this problem has been to
show the customer photographs of various hairstyles, and to modify a
selected style to fit the customer as well as possible. The success of
this technique depends heavily on the hairstylist's skill and experience
in adapting the selected style to the customer's physical characteristics,
and on the degree of communication between the customer and the
hairstylist. Frequently, however, the result does not meet the customer's
expectations, and there is a significant cost to the customer, the
hairstylist, or both.
Various attempts have been made to employ video technology to address this
problem. One approach is to use a computer "paint" program that permits
the capture of a video image in digitized form, and then the subsequent
manual editing of the digitized image to obtain a desired effect. The
digitized image is displayed by a computer on a video screen, and the user
edits the image by moving a pointer or cursor about the screen, and
selecting various commands to change the extent and color of the digitized
image. This method is tedious and time-consuming, however. Its use in a
hairstyling environment is totally impractical because of the experience
and training that would be required of the personnel operating the system.
Another approach that has been proposed employs two video cameras. One is
positioned on a picture of a model wearing a proposed hairstyle, and the
other is focused on the customer. The model's face is cut from the picture
of the hairstyle, which is then electronically overlaid with the customer
image. The result is a very artificial "cut and pasted" appearance, with
no allowance being made for size differences or different face shapes.
It will be appreciated from the foregoing that there is still a real need
for an improved technique for selecting hair styles and, more generally,
for adapting video images to be overlaid in such a manner that the
resulting composite image has a natural appearance. The present invention
fulfills this need.
SUMMARY OF THE INVENTION
The present invention resides in a video processing method and apparatus
that permits a selected hairstyle image to be overlaid on a customer image
in such a way that the resulting composite image has a natural appearance.
An important advantage of the invention is that it is relatively simple to
use and requires only a very short period of operator training.
Briefly, and in general terms, the method of the invention comprises the
steps of capturing the video image of at least one hairstyle, processing
the captured video image of the hairstyle to render it adaptable to a
variety of face sizes and shapes, capturing the video image of a
customer's face, and combining the video image of the hairstyle with the
video image of the customer's face, in such a manner that the hairstyle
image fits naturally about the face image. In the presently preferred
embodiment of the invention, the step of processing the hairstyle image
includes adding "fill" hair to the image in a region covering a range of
size and shape adjustability, and the step of combining the hairstyle
image and the face image includes deleting "fill" hair that would
otherwise cover the face image. In this way, the hairstyle image is
combined with the face image without unnatural gaps between the face and
the hairstyle and without unnaturally overlaying the hairstyle over the
face.
More specifically, for each pair of corresponding picture elements (pixels)
for the two images, the step of combining the images further includes the
steps of determining whether the hairstyle image pixel is either real
hair, "fill" hair, or neither, and determining whether the face image
pixel is within a previously defined face area. Based on these
determinations, the remaining steps of the method are: using the hairstyle
pixel in the combined image if the hairstyle pixel is of real hair, using
the hairstyle pixel in the combined image if the hairstyle pixel is of
"fill" hair and the face pixel is not within the face area, and using the
face pixel if the hairstyle pixel is of "fill" hair and the face pixel is
within the face area, or if the hairstyle pixel is of neither real nor
"fill" hair.
The method of the invention may also include the steps of moving the
hairstyle image with respect to the face image, and repeating the
combining step to be sure that the images have combined without unnatural
gaps or overlaying of hair over the face.
One aspect of the invention includes the steps of changing the color of the
hairstyle image, and repeating the combining step to be sure that the
images have combined without unnatural gaps or overlaying of hair over the
face.
Yet another aspect of the invention includes the further step of blending
the colors of the hair and face images in a region close to the line where
the hair and face images meet. More specifically, the blending step
includes selecting a face pixel immediately adjacent to a hair pixel;
then, for each primary color, determining the average primary color
intensity for the selected pixel and for all pixels adjacent to the
selected pixel, and replacing the primary color intensities indicative of
the color of the selected pixel with the averaged values of primary color
intensities. These steps are repeated to blend the color of all face
pixels adjacent to a hair pixel.
It will be appreciated from the foregoing that the present invention
represents a significant advance in the processing of video graphics data,
especially in the field of hair styling and related activities. Other
aspects and advantages of the invention will become apparent from the
following more detailed description, taken in conjunction with the
accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram showing how the information is stored and
utilized in accordance with the method of the invention;
FIG. 2 is a block diagram of the interconnected hardware required for
practicing the method of the present invention;
FIGS. 3a-3e are diagrammatic views of a customer face image, showing how
the outlines of the facial features are marked and a face area is defined;
FIGS. 4a and 4b together comprise a flow chart of the steps performed by
the computer in marking the customer face image;
FIGS. 5a and 5b are diagrammatic views illustrating how a selected
hairstyle image and a customer face image are logically combined;
FIG. 6 is a flow chart showing the steps performed in combining a hairstyle
image and a customer image;
FIGS. 7a-7d are flowcharts illustrating the color selection and color
conversion functions; and
FIGS. 8a-8c are flowcharts showing the steps performed in color blending a
customer image and a hairstyle image.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Overview of the method:
As shown in the drawings for purposes of illustration, the present
invention is concerned with techniques for adapting a selected hairstyle
to the video image of a customer's face without unnatural gaps or
unnatural placement of the hair over the face. Prior to the invention,
video processing of combined multiple images was a time-consuming task
requiring a highly experienced operator.
In accordance with the invention, multiple hairstyle images are
prepossessed to render them readily adaptable to a variety of face sizes
and shapes, and a novel technique is employed for combining a customer
image with a selected hairstyle image, without unnatural gaps or hair
placement, and without the need for highly experienced personnel. The
basic flow of information in this process is shown in FIG. 1.
The success of the invention depends on the preliminary steps of style
preparation and customer preparation. Style preparation is, of course,
handled without customer participation, and extreme care is taken to
prepare images of various hairstyles in such a way that they can be easily
adapted to a customer image. Customer preparation should also be performed
with great care, but need only be handled once for each customer. When the
preliminary steps are carefully performed, the final step of combining a
selected hairstyle image with a customer image is effected rapidly and
conveniently, such that multiple hairstyles can be adapted to a customer
in a matter of minutes.
Model hairstyles, as indicated by reference numeral 10 are captured in
video form, using a conventional color video camera, and digitized for
storage on a magnetic disk or some other convenient medium, as indicated
at 12. The hairstyles are typically modeled by different people, who are
placed in a standard and predefined relationship with the video camera
while their images are digitized and stored. Optionally, the same images
may be stored directly on video recording tape, as indicated at 14, for
later viewing by customers. Alternatively, still photographs of the models
may be provided to the customer for a selection of hairstyles. The next
step, at 16 is one of preprocessing each hairstyle image to make it more
easily adaptable to a customer image. This step has two basic components:
removing the image of the entire face, neck and shoulders of the model, to
leave only the hair, and then adding "fill" hair. A critical problem in
adapting a hairstyle for display on multiple customers is that one
hairstyle will simply not fit all customers. If the model's face is wider
than the customer's, there will be gaps between the hairstyle and the
customer's face when the hairstyle image is combined with the customer
image. If the model's face is significantly narrower than the customer's,
hair may extend across the customer's face to a greater extent than is
desired.
Both these problems are solved by the addition of "fill" hair in the
hairstyle image. "Fill" hair is a hair image of the same color and
apparent texture as the real hair, but it is an image that is synthesized
electronically. The "fill" hair is placed at the sides of the U-shaped
hairstyle, extending in over the face area. Later, when the hairstyle
image is combined with the customer face image, the "fill" hair fills any
gaps between the customer face and the real hair of the hairstyle image.
Thus, the hairstyle with "fill" can adapt to any face size narrower than
the model's. Adapting to a wider face is just as simple. For those cases
in which a model's face is thought to be narrow, some of the real hair in
the hairstyle can be converted to "fill" hair. As will become clear from
the later description, real hair in the style image is always displayed,
even if overlapping the face. "Fill" hair is displayed only if is not
overlapping the face.
As indicated at 18, the preprocessed hairstyles are saved on a storage
medium, such as a magnetic disk, for later use in a hairstyling salon.
First, however, the equally important process of customer preparation must
be performed. As indicated at 20 and 22, the customer image is captured on
magnetic disk in much the same way as the hairstyles. In the case of the
customer, however, the object is to capture the face image without any
hair. To this end, the customer wears a tight wig cap to draw most of her
hair to the rear while being recorded from a nearly frontal position. As
indicated at 24, the customer preparation steps are principally concerned
with defining the size and shape of the customer's face. As will be
explained in detail, the video image of the customer's face is defined by
marking a series of points on the outline of the face and ears. Once this
is done, the facial area is defined in the stored image. This is the area
on which "fill" hair will not be displayed, even if present in the
hairstyle image.
As indicated in block 26, the next step is to select a hairstyle from a
catalog of those stored on disk, and to combine the selected hairstyle
with the customer image. As already mentioned, the basic process employed
in the combination step is to display either the hair from the hairstyle
image or the face from the customer image, in accordance with a simple set
of logical rules. If, for any point on the combined image, there is real
hair from the hairstyle, then the real hair is displayed, even if over the
face. "Fill" hair is displayed only if not over the face area, and blank
areas of the hairstyle image are supplied from the face image.
As shown in block 28, the composite image obtained by combining the
hairstyle image and customer image may be adjusted in a variety of ways,
including adjustments in size, position, and color. There is also an
important option to "blend" the colors of the face with the adjacent hair.
This removes the sharp "staircase" edges sometimes associated with the
synthesis of graphics images. An optional step, shown at 30, is the
storing of the composite customer images on video tape. This helps the
customer in her ultimate selection of a hairstyle, since the tape may be
taken home for later viewing before a final decision is reached.
The hardware used in the method of the invention comprises conventional
modules that can be purchased "off-the-shelf" for use in this system. A
simplified hardware diagram is given in FIG. 2. The heart of the system is
a computer 32 compatible with the 16-bit Microsoft disk operating system
(MS-DOS), having 640k of random access memory and preferably a hard disk
drive as well as a floppy disk drive. The camera 34 and its power supply
36 are of conventional design, coupled to the computer by means of an
image capture board (ICB), the one used being manufactured by American
Telephone and Telegraph Company (AT & T). The computer is connected to a
composite color monitor 40 and, optionally, to a video cassette recorder
42.
The steps described briefly with reference to FIG. 1 will now be discussed
in more detail.
Hairstyle preparation
The preparation of hairstyles for inclusion in a catalog for selection by
customers can be performed without special-purpose software, since the
time taken in this process is not critical. Style preparation is basically
an "off-line" function, performed at some site other than the hairstyling
salon. The catalog of hairstyles will typically be compiled by the
supplier of the hairstyling system, with copies being supplied to
hairstylists who are to use the system.
In the presently preferred embodiment of the invention, style preparation
is performed by means of a graphics "paint" program called True-vision
Image Processing Software (TIPS), manufactured by AT & T, and specifically
intended for use with their image capture board. Basically, IPS and other
programs like it permit the user to edit captured graphics images, or to
create original graphics images in a computer. The editing function, which
the method of the invention uses here, allows the user to define, delete
or modify areas of the image, and to save the modified images for later
use. In effect, TIPS can be thought of as the graphics equivalent of a
powerful word processing program.
To capture the hairstyle, the model is placed in front of a brown
background and her position is adjusted until her head in the
corresponding video image fills predefined image oval. Then the image is
captured and digitized for subsequent processing. A preliminary step is to
replace all picture elements (pixels) in the image that are absolute white
in color with a near-white color. The absolute white color is used to
indicate a gap in the hairstyle when the image is later stored, so it is
important to eliminate this color from the actual hair. The specific
details of the next following steps will depend on the specific editing
program being used, but the TIPS terminology is used here by way of
example. It will be understood, however, that it is the function being
performed that is important to the invention.
The "touch-up" function of TIPS is used to place a white border around the
outside of the hair, replacing the blue background. The touch-up function
provides a simulated "magnifying glass" effect to permit editing of the
image on a pixel-by-pixel basis. The same function is then used to edit
out, or remove, the face of the model, including the ears and the flesh
tones in the forehead bangs of the hairstyle. At the end of this process,
there should be a white border completely around the hairstyle. The next
step is to build a mask to protect the hairstyle image from the subsequent
steps. In TIPS this is done using the functions "mask" from a special
effects menu, then "clear" to clear any prior selection of a mask, and
"build" to build the new mask. Building the mask is completed by selecting
the "fill" tool twice and then the "boundary" fill function. The system
requests that a cursor be placed over the white boundary, and then over
the hairstyle area to be protected by the mask, which will then appear as
a purple-colored area on the screen.
To use the mask that has been created, the "mask" function is again
selected (twice), and then the "use" function. Again the "boundary" and
"fill" functions are selected to choose the white border area. The result
is to fill the entire area except the masked hairstyle with a solid white
color.
The next step is to set an identifying bit in each pixel making up the
hairstyle area. This bit is subsequently used to identify real hair in the
hairstyle image. In TIPS the identifying bit is set by selecting the
special effect "video," which first blacks out the screen, then selecting
the sub-menu options "sub" and "video only." Then the tool option
"rectangle" is selected twice, and the "filled" option from its sub-menu.
When exit is made to the image display, the entire picture area is filled
in by defining its upper-left and lower-right corners, leaving the
background white once again and the hair area filled with black. When
return is made to the TIPS menu and the "video" option is deselected, the
hair should reappear on the screen against its now white background.
The final and most important step is to add some extra or "fill" hair to
the style image. In TIPS the "tile" function is used to match the "fill"
hair as closely as possible to the real hair of the hairstyle. The "tile"
function permits the user to select an area of the image that can be
copied in color during a subsequent "draw" operation. The user can select
a large cursor from a cursor menu, and then select the "draw" tool. "Fill"
hair is added using the selected cursor. The "fill" hair is added to the
inside region of the hairstyle, to permit it to be more easily adapted to
wider customer faces. If necessary, real hair can be changed to "fill"
hair using the appropriate editing functions.
As previously noted, the foregoing description is specific to the use of
the TIPS package for editing the style image. Any similar editing software
could be used for this purpose. The important aspects of processing the
style image are that the background is totally removed, i.e. converted to
pure white, that "fill" hair is added to the image, and that pixels
representing real hair are in some manner tagged for later identification.
The captured hairstyle is complete at this point except for aspects
relating to its color. A "COLOR" program analyzes the pixels making up the
total hairstyle and determines the average red, green and blue content of
the hairstyle. These are stored with the hairstyle to properly
characterize its color. Other parameters stored with the hairstyle give
its style number and name, and characterize it by face shape, length, and
cut type. A color conversion program is available for use at the time the
hairstyle is combined with the customer image. Use of this color
conversion program depends on the existence of recorded parameters to
effect the color changes, together with a copy of the hairstyle in its
natural color, prepared as part of the style preparation procedure.
Ideally, the color of the hairstyle should be convertable to any of
twenty-eight colors, including eight ash colors, eight gold colors, six
red colors and six silver colors. Since hair color conversion is a process
that defies perfection, it is best performed during style preparation,
when trial and error may be used without waste of customer time. The
hairstyle may then be stored in as many hair color versions as possible in
the style catalog.
Customer Preparation
After style preparation, customer preparation is the second preliminary
step for rapid adaptation of selected hairstyles to a customer's face. In
customer preparation, the customer wears a tight wig cap to hold her own
her hair toward the back of her head, and takes a position facing slightly
to one side of center. As in style preparation, an oval shape displayed on
the color video monitor is used to align the customer's head properly.
Then the image is captured and processed to properly record the face size
and shape.
The procedure includes prompts to the operator to record the positions of
prescribed points on the customer's facial image. In the first step, a
horizontal cursor line appears near the top of the display. The operator
positions this line at the hairline on the customer's forehead, as
indicated at A in FIG. 3a. For convenience, a computer input device known
as a mouse is best employed for movement of the cursor. The mouse is moved
across a flat surface and this motion is converted into cursor movement.
When the operator wishes to indicate that the appropriate position has
been reached, a button is actuated on the side of the mouse. Accurate
control is thereby achieved with only one hand, and the actuation of
special keys on the keyboard is not needed.
Similar steps are performed to mark the lower extremity of the chin, and
the right and left sides of the face, not including the ears, as indicated
at B, C, and D, respectively. As shown in FIG. 3b, in the next series of
steps, four successive diagonal lines are displayed on the face image, and
the operator must position the cursor line at the points of intersection
of these diagonal lines and the diagonal extremities of the face, as shown
at E, F, G and H. As each point is marked, it is displayed as an X on the
screen.
These first eight points defining the facial shape are shown in FIG. 3c by
the numerals 1-8 in circles. Next, as shown in FIG. 3d, successive points
9-28 must be marked in sequence by the operator. These are the upper and
lower extremities of the right ear, five additional points defining the
right ear contour, the upper and lower extremities of the left ear, five
additional points defining the left ear, and three points on each side
defining the neck and shoulder outline.
After all twenty-eight points on the extremity of the face have been
defined, the computer sets an identifying bit in every pixel of the face
area, as defined by the shaded area of FIG. 3e. Basically, the face area
is defined for this purpose as extending from about the middle of the
forehead and down across the ears, neck and visible portions of the
shoulders. As mentioned earlier, this defined face area is used to
determine whether or not to display "fill" hair. If "fill" hair falls on
the face area, it is not displayed real hair is displayed, whether in the
face area or not.
FIGS. 4a and 4b together define a flow chart of the functions performed by
the computer in customer image preparation. After entry to the procedure,
at 50, a brief operator message is displayed (52), and a subroutine 54 is
entered to make sure that the identifying bits in all of the pixels are
off. Then, in accordance with a series of decision blocks 56-60, it is
determined which of the twenty-eight identifying points is being addressed
on the face. For the first two points, a horizontal cursor is selected and
displayed, as shown at 61. For the third and fourth points, a v | | |
