Technical Field

The present invention relates to a device and method for simulating or predicting redistribution of the volume of a human body due to cutaneous pressure exerted by a garment worn by the human.

Background

Measuring and modeling the body shape of a person can be desirable for many reasons. The field of measuring and modeling a person's body for the ready-to-wear industry is increasingly popular. Methods have been presented that aims at predicting what garment fits a certain body, in terms of tailor's pattern, size etc.

WO0101354 shows a method and apparatus for creating computer graphical representations of a person in different poses, and how they can be used to display a variety of different garments on a model individual, and further how a garment alters the silhouette and appearance of a model individual. Using a multitude of CCD cameras, and structured light, images of the persons various limbs are obtained. Said images, plus the person's weight are applied to a stored generic model avatar which has been created from knowledge of how an unclothed generic person varies with weight. WO0101354 discloses a garment processing system that can model how a garment alters the shape and appearance of the wearer's silhouette. WO0101354 refers to providers of 3D laser scanning and discards the use of them, claiming it can only obtain data indicative of the surface of an individual in a single pose. Instead WO0101354 relies on mapping photo data onto weight adjusted generic avatars, which means that the less "generic" the actual persons body shape is, the less accurate the result will be.

This is a most unfortunate drawback, as it is always more difficult to find well- fitting retail clothes if you have an untypical body constitution, be it excess overweight, particularly long, short or uneven limbs, a very large chest or hyphosis.

US5530652 assigned to Strauss & Co, discloses a system that can be used to measure a person. A person stands on a rotating platform while measurements are taken.

US5530652 suggests that a laser imaging system may be used, but the preferred embodiment comprises a CCD camera. The camera captures images at different predetermined angles of rotation, e.g. at every ten or twenty degrees.

Used in a retail store the customer is expected to put on a tight-fitting suit and then to stand on the rotating platform legs. In order to get useful images the person is suggested to have a cone between his or her legs, so as to spread them. When the rotating platform has rotated through a full 360° turn, a computer creates a

computerized wire-frame image, that is, a 2D image of how a 3D model of the person would look. A similar procedure is performed on pants. Air-inflated, the pants are being rotated, images are captured and a wire frame is produced. The wire-frames may be matched against each other to determine the best pattern of a pair of pants, such as jeans, for this customer.

The system disclosed in US5530652 is marred by drawbacks. First of all, having to stand on a rotating turntable in a leotard-like costume with what appears to be a large cone between your legs must be considered equally humiliating for both genders. Secondly, although the person being modeled will find a pair of jeans that fits when he stands still, legs stretched and spread, there is no way of knowing whether he will be comfortable sitting, squatting or striding. Thirdly, the way especially jeans are worn tight, this method will not do. Cotton worn tight yield in time. A traditional pair of Levi's © jeans, much like a pair of shoes, must be broken in. Therefore a person may very well prefer a pair of jeans that are not comfortable to begin with, knowing that certain areas of the jeans will yield. For that matter, it will not work for stretch jeans either, as they will not react to a fan as they will to being worn properly. Fourthly, although the solution is described as a Three-Dimensional Measurement System, it is in fact rather 2D to character. Pictures are taken at discrete angles, there is no real 3D modeling, but rather comparison between frame-wire images - in effect comparing 2D to 2D. This makes the method unreliable.

US2009137894 aims at solving the problem of finding a bra with a proper fit. A woman sits in a large transparent tank, high enough to allow her full torso. The tank is filled with salt water of salinity adjusted according to the woman's BMI (Body Mass

Index, i.e. body weight divided by height squared). The buoyancy of each breast then accomplishes a lift. A scanner, referred to as a 3D scanner, scans the woman's upper torso while moving in circles along rails on the outside of the large tank. The scanning data is used to calculate size, volume and location of each breast and the data is used to select a properly fitting bra which holds and supports the breasts in the position they were when scanned. The device and method according to US2009137894 owns several severe drawbacks.

Firstly, very few women are comfortable with being asked for their BMI, which is essentially the same thing as being asked to what degree you are overweight. Secondly being asked to skinny-dip in a transparent tank, and to remain completely immovable, for the duration of a scanning procedure is not aimed at making people less uncomfortable. From a CRM (Customer Relations Management) perspective, therefore, the method of US2009137894 shows several disadvantages.

Thirdly, the management of the contraption disclosed in US2009137894 demands a facility that easily blends at least 1000 liters of new saline solution to order, for every new client, and this saline solution must be pumped in and out of the tank, and then be disposed of, for hygiene and CRM reasons.

Fourthly, a woman's body changes over time, not least during pregnancy and breastfeeding. These are also periods where a woman's need for new, well- fitting bras is greater than ever. And during these exact periods, using BMI to calculate the proper saline concentration for the proper level of breast buoyancy is less accurate than ever. Although a breast usually contains mostly fatty tissue, during pregnancy and especially nursing, the fat/water ratio changes dramatically, and with it the breast buoyancy.

Therefore, the method is unreliable, and may produce inaccurate results.

There is certainly a need for devices and methods that not only facilitates garment fitting, but also truly delivers a better end result than actually trying on every available garment. Summary

The object of the present invention is to obviate at least some of the above disadvantages and provide improved methods, apparatuses and computer media products avoiding the above mentioned drawbacks.

Particularly it is an objective of the present invention to offer a solution to the problem of how to predict how a certain given garment will fit a certain body in such a way that the natural body shape is moulded, i.e. temporarily altered by the garment, so as to assume a desirable figure. A further objective is to enable to take into account a garments tendency to yield to stress over time, so that it is possible to simulate and predict how the shape of a body looks when wearing a well-worn garment. It is further an objective to offer accurate real-time predictions of how a body wearing a garment changes shape as the body takes various postures. It is a further objective to offer devices and methods that can do all the above quickly, conveniently, hygienically and in a way that is not humiliating. It is further an objective to offer solutions that do not require bulky equipment, a large dedicated area, specialized drainage or fan systems. A first aspect of the invention is a method for simulating or predicting redistribution of the volume of a human body due to cutaneous pressure exerted by a garment worn by the human. The method comprises the steps of capturing real-time 3D range data from at least a portion of the body to be covered by the garment with a TOF sensor; modeling a before-avatar of the captured portion based on the results of the TOF range data; simulating volume redistribution by generating an after-avatar as a function of the before-avatar and the garment's tensile stress properties exerted on the body; and rendering a simulated image of how the body would appear with the garment worn.

According to certain embodiments of the present invention, the method comprises the further step of applying gesture control in order to predict the body's freedom of movement in the garment.

According to certain embodiments, the tensile stress properties of the garment comprise both short-term elastic and long-term plastic properties.

A second aspect of the invention is a device adapted and configured to perform the method of the first aspect of the invention. The device for simulating or predicting redistribution of the volume of a human body due to cutaneous pressure exerted by a garment worn by the human comprises a TOF sensor, a processing unit adapted and configured to receive range data from the TOF sensor, use the range data to model avatars and to simulate the effects of garment pressure, a memory unit, an input/output unit comprising a rendering device adapted and configured to render the results of simulations in 2D or in 3D. The device is further adapted and configured to capture realtime 3D range data from at least a portion of the body to be covered by the garment with a TOF sensor; to model a before-avatar of the captured portion based on the results of the TOF range data; to simulate volume redistribution by generating an after-avatar as a function of the before-avatar and the garment's tensile stress properties exerted on the body; and to render a simulated image of how the body would appear with the garment worn.

According to certain embodiments the tensile stress properties of the garment comprise both short-term elastic and long-term plastic properties.

According to certain embodiments the memory unit is further adapted and configured to comprise a garment information database, and the rendering device is adapted and configured to retrieve garment information from the database and render it, thereby enabling an operator of the device to select a set of garments comprising at least one garment as input to simulation by the processing unit. According to certain embodiments, the device is further adapted and configured to apply gesture control in order to predict the body's freedom of movement in the garment.

A third aspect of the present invention is a computer program product comprising program code means stored on a computer readable medium for performing the method of the first aspect of the present invention when said product is run on a computer.

Brief Description of the Drawings

In order to explain the invention in more detail an embodiment of the present invention will be described in detail below, reference being made to the accompanying drawings, in which

Figure 1 is a schematic view of a device according to an embodiment of the present invention.

Figure 2 is a flow chart illustrating the method steps of an embodiment of the present invention.

Detailed Description

Though in the field of modeling and visualizing peoples bodies for the purpose of fitting clothes, previous art as described above fail to acknowledge the positive aspects of garments as body shapers that redistribute a body's volume as a desired effect. Though WO0101354 sets out to predict the shape of a person wearing a certain garment, the proposed solution is to model the outer garment silhouette, not body shape.

Though jeans are to a degree the corsets of our time, US5530652 fails to identify a need to take the body shaping properties into account. Using the US5530652 device will not help a person find tight jeans that will fit like the proverbial glove in time, or a person who is looking for a pair of jeans that will flatten the tummy, or mould a J-Lo butt.

Lastly, though US2009137894 does acknowledge and promote redistributed body volume, they do it in a peculiar and most unwieldy way. In US2009137894, there is just one right shape of a body - the "natural shape", and obtaining that shape is merely a means to produce a garment. The present invention on the contrary uses the garment as a means to produce one of several body shapes, which may have never been a "natural shape" at all. Moreover, every woman knows that while a bra may be perfectly comfortable and possible to wear under a t-shirt, it may be useless under a particular dress: a certain top requires a bra that flattens the bust, another demands the breasts to be lifted, to give them a certain cleavage or shaping them more or less pointed. And while "perky" may be a desirable look, "buoyant" is not - it is doubtful if any woman would want her body shaped with the US2009137894 device. This goes to show that a person skilled in the art does not obviously identify the problem of how to predict and simulate a desired redistribution of body volume caused by a certain garment.

Therefore we are suggesting a system and a method that measures a body of a person, models it in real-time, applies information on how certain areas of the body react to different kinds of garment-induced pressure, contains garment models that incorporates information of how the garment yields to or resists deformation under tensile stress in various directions, and over time.

A system according to embodiments of the present invention will now be described in relation to figure 1, which shows a device 100 for simulating or predicting redistribution of the volume of a human body 10 due to cutaneous pressure exerted by a garment 20 worn by the human. The device 100 comprises a TOF sensor 110, a processing unit 120, a memory unit 130, and an input/output (IO) unit 140 comprising a rendering device 150.

The TOF sensor 110 comprises an illumination unit, and a sensor arrangement comprising a pixel sensor plane. The illumination unit is adapted and configured to emit a detection signal comprising a certain wavelength band, with a certain modulation scheme. Each sensor pixel in the pixel sensor plane is adapted to receive the reflected signal. The sensor arrangement is adapted and configured to simultaneously convert variations in the scattered, or reflected, signal into a snapshot image frame comprising range information using properties of the signal as reference.

The processing unit 120 is adapted and configured to receive instant range data from the TOF sensor, use the range data to model avatars and to simulate the effects of garment pressure. The processing unit 120 is connected to a memory unit 130, adapted and configured for storing predefined data, as well as intermediate storing and storing of final results. The memory unit may further comprise a garment information database

135. Tensile stress properties of garments comprising both short-term elastic and long- term plastic properties may be stored in the database 135.

The IO unit 140 is adapted and configured to interact with a human operator 30 of the device 100, and comprises a rendering device 150 adapted and configured to render the results of processing unit simulations in 2D or in 3D, such as holographic rendering. The rendering device may also comprise CNC type automatic mode milling functionality or other prototyping machines that are able to build hardware models.

Due to the TOF unit's snap-shot character, no time-consuming scanning or bulky support rails are needed. With appropriate data and software, one TOF unit is enough to generate a digital model. However, in certain embodiments, more than one TOF unit is comprised, in order to additionally enhance the accuracy of the simulations. Because the TOF unit capturing is instant, and because the TOF unit is significantly smaller that e.g. scanning systems comprising moveable parts, there is no need to confine the client to a dedicated area during the process. Therefore, the system can be made compact, inconspicuous and easy to move between different premises.

In relation to figure 2 and also figure 1 is described a method for simulating or predicting redistribution of the volume of a human body 10 due to cutaneous pressure exerted by a garment 20 worn by the human to whom the body belongs. The method comprises the steps accounted for below.

In a capturing step 220 real-time 3D range data is being captured by the TOF sensor 110, from at least a portion of the body 10. In certain embodiments more than one portion of the body 10 may be 3D captured sequentially with the same 3D imaging unit. Other embodiments comprise more than one 3D imaging unit, thereby allowing the system to receive simultaneously range information through reflected signals from all over the body 10. Due to fringe effects, such as bulging of body tissue along the hems of a tight garment, certain embodiments capture a portion of the body large enough to comprise these fringe areas as well, and not just the area covered by the garment 20.

In a modeling step 240 the processing unit 120 is modeling a before-avatar of the captured portion based on the results of the TOF range data.

In a simulating step 260, volume redistribution is simulated by generating an after-avatar as a function of the before-avatar and the garment's 20 tensile stress properties exerted on the body 10. Generic information on how different parts of a human body react to cutaneous pressure may be pre-stored in the memory unit 130, and applied to the surface of the before-avatar. For further reference, elasticity profile is defined as a function describing resistance to cutaneous pressure in any relevant point of the surface of a person's body. Based on relative body measurements retrieved from the before-avatar, such as e.g. hight/hipmeasure ratio and joint positions, a person's individual elasticity profile can be relatively accurately determined, without

intimidating questions or invasive caliper measurements. This is especially true for the female chest area. The after-avatar is further generated as a function of tensile stress properties of a specified garment. The tensile stress properties of a garment decides to what extent a garment will fold, drape, stretch, or bulge under stretch among other things. It may also decide what tension rips the fabric or the seams, as well as where it rips. Further the tensile stress properties of a garment may comprise both short-term elastic and long- term plastic properties. This is advantageous both because it is possible to detect and avoid garments whose visual appearance will deteriorate, and because it is possible to se exactly how a pair of un- washed unworn jeans will fit and look when they are properly broken in.

In a rendering step 280 a simulated image 40 is rendered of how the body 10 would appear with the garment 20 worn. The simulated image may be a wire frame image created from the after-avatar.

In certain embodiments, the simulation step 260 may employ a default garment pre-stored in the memory unit 130. In other embodiments however, the client may select a garment 20 for simulation. This may entail pointing at one of many garments rendered on a screen comprised in the I/O unit. The client may also choose to select a set of garments for simulation. Every garment is associated with its specific tensile stress properties. The step 260 of simulating volume redistribution is then being repeated for every garment comprised in a set of garments selected, by default or by input. Several after avatars may then be generated, and several images 40' may be rendered.

In certain embodiments, the method comprises the step of receiving input specifying a target-avatar. The client may be offered to select a desired body shape from suggested body shapes rendered on a display. The rendered body shapes may be rendered based on what is possible considering the before-avatar. Each of these body shapes may be associated with a target-avatar. In these embodiments, each after-avatar generated during simulation may be stored intermediately in the storage. In a selecting step, after-avatars are compared one by one with the target-avatar. The after-avatar that deviates the least from the target avatar is selected.

The rendering step 280 may comprise rendering an image 40' of each generated after-avatar. If an after-avatar is selected, the rendering step may comprise rendering an image of the selected after-avatar. The output may indicate what garment has properties such that the shape of the selected after-avatar may be attained.

In certain embodiments, the method comprises the further step of applying gesture control, in order to predict the body's 10 freedom of movement in the garment 20. For instance, this enables predicting whether a specified garment 20 may break as the human body 10 takes a certain posture. It is possible to simulate the outcome of a set of predefined common postures, such as for instance sitting down, bending over to reach the floor, climbing a high step, lifting arms over head. In certain embodiments the 3D image units are capable of rendering real-time 3D video data of the client moving about, this to enable the client to see him or herself virtually moving and gesturing in a certain outfit.

Although the three documents WO0101354, US5530652 and US2009137894 concern related art, neither of them realizes the merit of a TOF camera, this even though at least two of them implicitly relates to "3D laser measurements".

For completeness sake, documents US2001026272, US2009222127 and US2009138377, disclose garment design/clothes fitting systems, especially for flexible fabric, and methods directed towards visualizing a desired level of strain or pressure through color indicators or the like. These latter three documents are not directed towards simulating or predicting redistribution of the volume of a human body, and do not offer solutions to the problems towards which this patent application is directed. However, US2001026272 asserts that it is desirable to perform a property analysis on the model in order to simulate skin or bone deformation, but fails to disclose how such an analysis could be done. According to the teachings of US2001026272, 3D methods, notably such as is disclosed in US2010111370, fail to accurately handle clothing that obscures underlying body shape. Further US2001026272 discards 3D scanning methods as complex, expensive and too specialized and instead advocates low-dimensional parametric modeling as superior to 3D based methods.

This goes to show that the act of implementing a snapshot type of range data imaging system, such as a TOF camera, is not obvious to a person skilled in the closest art of redistribution body volume, and further, that teachings in other arts such as garment design actually teaches away from such methods.

Further advantages compared to previous art is that this method is quick; it does not expose the client in a humiliating way; there is no need for uncomfortable questions about body weight; it does not require a lot of space or a dedicated area; being true real-time 3D, the prediction is accurate, as opposed to 2D simulations produced from a generic model.