KR20100050052A

KR20100050052A - Virtual glasses wearing method

Info

Publication number: KR20100050052A
Application number: KR1020080109141A
Authority: KR
Inventors: 김영준
Original assignee: 김영준
Priority date: 2008-11-05
Filing date: 2008-11-05
Publication date: 2010-05-13

Abstract

PURPOSE: A virtual glasses wearing method is provided to experience wearing virtual glasses through a simple procedure to click several reference points. CONSTITUTION: A facial image of a glasses wearer is inputted. Several position reference points(100,110,120,130) are appointed to the inputted facial image as described above. According to the position reference points, the rotated value of the facial image to xyz axis, the size of glasses image, and the position of the glasses are calculated. According to the value calculated as described above, a transformed glasses image is created. The created glasses image and the facial image of the wearer are composed.

Description

Virtual Glasses Wearing Method {Virtual Glasses Wearing Method}

The present invention relates to a method for virtually wearing glasses on a face image of a wearer by using a computer. In particular, the glasses image is deformed according to the rotation value of the face image of the wearer. The present invention relates to a virtual glasses wearing method for generating a virtual wearing image to directly select glasses suitable for the wearer without wearing the glasses directly.

Recently, eyeglasses have become an important fashion accessory that expresses not only the original medical function that protects the eyesight of the user but also the taste and personality of the eyewear wearer.

In the past, in order to purchase such glasses, most users went to the store and tried to buy glasses of various types, and then chose glasses to buy. As e-commerce increases, the glasses, like other fashion products, are sold on the Internet site. The number of users who purchase and view the images they provide is increasing. Especially, sunglasses, goggles, and goggles, which are not directly related to eyesight, do not need the help of specialists such as doctors or opticians. have.

However, when purchasing glasses only by looking at the glasses image through the Internet, the satisfaction is lower than when actually wearing the glasses, and using the computer image synthesis technology to create a virtual wearing glasses to increase the user's satisfaction Various virtual wearing methods have emerged.

In the related art, Korean Patent Registration No. 10-0365664 proposes a method of simulating a state similar to the actual state of wearing glasses by adding pre-inputted prescription values of subjects to wear and lens and eyeglass frame information, and No. 61-80222. In the publication, a portrait of a person is photographed with a video camera, the spectacle wearing area is designated on the person, and the frame is superimposed on the designated area, or Korean Patent Registration No. 10-0386962 shows a face image photographed from various angles. The standardized input and standardization of the glasses image and the actual dimensions of the glasses were inputted and standardized to synthesize the glasses image on the face image, and presented a virtual image at various angles.

However, in the former case, the optician is selected as the main object of the proposal, and the wearer visits the visitor in person to generate the wearer information.In addition to the shape information, the glasses information to be worn includes detailed information such as the material, lens type, frequency, thickness, and weight of the frame. There is no need to provide a lens from the medical side because it requires management, there is a problem that is difficult to apply to the store of the e-commerce method that the user must operate the system directly.

In the latter case, first of all, in terms of the user, firstly, it is difficult to apply to electronic commerce because it is necessary to have an expensive system capable of photographing glasses wearers in various directions or move to a separate place equipped with the system. The second is the process of receiving and processing only the front picture as the network upload method, but in this case, the process of inputting the face dimension information of the user is complicated, and the face dimension information and the face shape are analyzed automatically by the system. The method of extracting information also has a problem in that it is difficult to purchase glasses that are actually satisfactory because the accuracy is significantly lowered due to various photo backgrounds when applied to the photograph taken by the user. Third, the method of changing, applying and purchasing only glasses by continuously utilizing the once photographed image is not realistic given the fact that the preferred glasses are changed according to their clothes, hairstyles, etc. as they are already established as fashion props. Fourth, the size of the captured image is stored in each step and the original is damaged, and the loss of the image size can occur.

In terms of service provider, it is necessary to generate three-dimensional shape information of glasses for synthesizing glasses images from various angles. This process takes a lot of time and money, and stores the generated three-dimensional results in stages. As a result, storage space is wasted. In addition to this, there are many problems such as having to visit the wearer at least once and photographing using professional equipment.

In addition, Korean Patent Registration No. 10-2004-0097349 proposes a method of providing a virtual wearing image by generating a face of a wearer as a 3D model and synthesizing various glasses generated as a 3D model, but provided in 2D. A high performance computer is required in the implementation stage of analyzing the face photo and generating a 3D model, and the obtained result is also a deformation of the provided face of the wearer.

Accordingly, an object of the present invention is to provide a virtual glasses wearing method for the user to easily and variously create a variety of optimal eyewear pattern without any additional professional knowledge or information in order to solve the above problems.

According to an aspect of the present invention, a first step of inputting a face image of a wearer, a second step of designating a position reference point in the input face image, and a face size and an eye position according to the position reference point And a third step of calculating a rotation angle, a fourth step of generating an image of the glasses modified according to the value calculated in the third step, and a fifth step of synthesizing the image of the glasses generated on the image of the wearer. Characterized in that it comprises a.

According to the present invention, a user of the present invention who wants to purchase glasses can experience the virtual wearing function of the glasses through a simple process of selecting and inputting a picture of the wearer and clicking only a few reference points, Since the result can be virtually worn by using the three-dimensional rotation function of the glasses even if the head is not exactly facing, the user does not have to photograph a separate image for use of the present invention. The invention can be applied.

In addition, the image of the provided face is used as it is, and the glasses also transform and synthesize 2D image information, so the computation amount is smaller than that of the 3D modeled glasses composition, so that the user can directly perform the deformation and compositing functions of the glasses on the user's computer. have. In this case, there is no need to transmit the wearer's image to the server, which protects the user's personal information, and the server transmission process of the wearer's face image, which is relatively large compared to the glasses image, is omitted, thereby reducing the execution time. Can be.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

First, the user executes a virtual wearing program that is downloaded or pre-installed through the Internet, and inputs a face image of the wearer into the program. In this process, the server transmission function of the face image may be omitted to protect personal information and prevent execution delay of the program.

A reference point is then input to determine the position, size and rotation of the glasses on the wearer's face image. In general, when taking a face image, the head is often turned to the left or right, or bowed up and down to obtain a better quality, so even a front view may cause a considerable amount of rotational error. Therefore, in order to obtain a realistic fit, the glasses image should be rotated or moved according to the input face image. For this purpose, the user can calculate the rotation angle to know how tilted the wearer's face image is from the front. Specify a reference point, such as 1. These reference points are assigned to the eyes and ears, for example, as the center points 100 and 110 of each eye, and the uppermost points 120 and 130 of the right and left ears respectively touching the face.

In FIG. 2, the vertical distance 230 of the reference point 100 of the right eye and the reference point 120 of the right ear input by the user, and the vertical distance 240 of the reference point 110 of the left eye and the reference point 130 of the left ear 130. Is compared with a statistical mean to calculate the upper and lower lift (x-axis rotation value, hereinafter x) of the face, and the difference between the reference point 100 of the right eye 100 and the reference point 120 of the right ear 200 and the left By comparing the difference 210 between the reference point 110 of the eye and the reference point 130 of the left ear, the degree of left and right rotation of the face (y-axis rotation value, hereinafter referred to as y) is calculated, and the angle of the line connecting the two eyes is calculated. The measurement calculates the inclination (z-axis rotation value, hereinafter referred to as z) 250 of the face. This value can also be calculated by measuring the angle of the line connecting the two ears.

The size t of the glasses image to be synthesized is determined by using any one of the distance 220 of the reference points 100 and 110 of each eye and the distance 260 of the reference points 120 and 130 of each ear, or by using both values. .

The rotation values (x, y, z) of each axis, the size (t) of the glasses image, and the position of the glasses in the face image calculated through the above process are stored together with the image file information. Even if another pair of glasses is selected using the image file, it is not necessary to repeat the process of inputting the reference point again.

The virtual wearing glasses program generates a modified glasses image by using the rotation axis or the rotation reference point information of the glasses according to the size, the rotation value (x, y, z) of the glasses.

The rotation axis or rotation reference point information of the glasses is provided by the server administrator.In order to generate more accurate glasses deformation image during compositing, the information about the wearing point where both eyes and ears are located is separately stored for each glasses, and the effect of rotation is used. In order to increase the information about the rotation reference line or rotation reference point of each of the x, y, z axis is also stored for each glasses.

For reference, the user may select the glasses in any process.

In FIG. 3, when the glasses image 300 is completed through the above processes, the glasses image is automatically synthesized on the input face image. In this case, if the user does not like the automatic synthesis, the user can directly adjust the position and rotation value of the glasses directly. For example, the mouse can directly move or rotate the glasses image to a desired position.

And when the user selects the other glasses, as shown in Figure 2, the size of the glasses and the position and rotation value information on the face is stored, so the user can get a composite picture applied with the new glasses without a separate reference point input process Can be.

Meanwhile, in order to increase the wearing comfort of the synthesized glasses image, the color tone of the glasses may be changed according to the background of the input image. For example, by analyzing the color tone information of the periphery where the glasses are to be worn in the input image, and calculating the position of the light source. When the glasses image is combined with the face image, the lens tint is changed, and the light is reflected off the lens or creates shadows.

In addition, according to the y value, all or part of the pre-stored leg image 310 may be further synthesized on the right or left side to increase the virtual wearing quality of the glasses. It exposes the image of the glasses legs of the face exposed a lot according to the direction of the y-axis rotation value, the distance 200, 230 and the left eye 110 between the right eye 100 and the right ear 120 of FIG. The length of the glasses legs exposed based on the frame of the glasses may be adjusted using the distances 210 and 240 between the left ear 130 and the left ear 130.

1 is a view showing a wearer's face image and the reference point input by the user in accordance with the present invention.

FIG. 2 is a diagram illustrating the calculation principle of the up / down lifting angle (x-axis rotation value), the left and right turning degree (y-axis rotation value), and the tilt of the face (z-axis rotation value) of the face image using the reference point of FIG. 1. .

FIG. 3 is a view in which the image of the glasses moved and rotated to a value calculated according to the principle of FIG. 2 is combined with the face image of FIG. 1. FIG.

Claims

A first step of inputting a face image of the wearer, a second step of designating a position reference point in the input face image, a rotation value of which the face image is rotated with respect to x, y, and z axes according to the position reference point; A third step of calculating the size of the glasses image and the position of the glasses, a fourth step of generating an image of the glasses modified according to the value calculated in the third step, and an image of the glasses created in the image of the wearer Virtual glasses wearing method comprising the fifth step of synthesizing.

The method of claim 1, wherein the location reference point is both eyes and ears.

The method of claim 1, wherein the fourth step comprises changing the color tone of the spectacle lens according to the background of the input face image.

The method of claim 1, wherein the fifth step includes adding an image of right or left leg of the glasses according to the y-axis value.

The method of claim 4, further comprising displaying only part or all of the image of the leg according to the y-axis value.