KR102538783B1 - Apparatus and method for generating a hairless head person image for a virtual hair styling experience - Google Patents

Abstract

The present invention relates to an apparatus for generating a bald person image and a method for generating a bald person image by the same. The method for generating a bald person image comprises: a step of dividing a person image, and extracting a background mask with a person object region removed, and filling a part from which the person object region is removed in the background mask by in-painting, and generating a background image; a step of extracting a face region in the person image, and generating a face image, and inputting the face image into an artificial neural network generating a head shape in the face region, and generating a bald head shape image; a step of synthesizing the background image and the bald head shape image, and generating a bald head shape background image; a step of dividing the person image, and extracting a non-hair mask from which a hair region is removed; and a step of synthesizing the non-hair mask and the bald head shape background image, and generating a bald person image. Therefore, a natural-looking person image, which is made by removing only hair from the face image of a person with hair, can be generated.

Description

Apparatus and method for generating a bald person image for virtual hair styling experience

The present specification relates to an apparatus and method for generating an image of a bald person for a virtual hair styling experience.

This invention is the result of an application filed with the support of 'TIPS (Startup Growth Technology Development Project, TIPS Project - S3200667)'.

Recently, in the field of beauty, a virtual hair styling experience service is provided in which customers experience various hairstyles they want in advance and decide which hairstyle to apply. The virtual hair styling experience service is a service that provides the experiencer with the same experience as having previously applied a hairstyle by synthesizing or superimposing a hair image generated virtually on a face photo of the experiencer and outputting the result. As the hair object image, a computer graphic design was used, and a virtual experience was provided by overlapping the designed hair object image on the face image of the experiencer and partially adjusting the size or shape as needed. .

The above description is only intended to help understand the background of the technical ideas of the present invention, and therefore, it cannot be understood as the prior art known to those skilled in the art.

However, the virtual hair image superimposed on the experiencer's existing hair does not naturally overlap with the experiencer's head or face, causing awkwardness, and the existing hair protrudes under the virtual hair image, resulting in a very low experience satisfaction. .

In addition, although the virtual hair image may be superimposed after removing the existing hair from the experiencer's face image, there is a problem in that the experiencer's experience satisfaction is lowered because the face image from which the existing hair is removed is unnatural.

The present specification is intended to solve the above problems, and an embodiment of the present specification aims to create a natural person image in which only hair is removed from a face image of a person with hair.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

The present specification proposes a method for generating an image of a person with a bald head using an apparatus for generating an image of a person with a bald head. The bald-headed person image generation method includes the steps of extracting a background mask from which a person object area is removed by performing segmentation on a person image, and filling in a portion from which the person object area is removed from the background mask with inpainting to generate a background image. ; generating a face image by extracting a face region from the person image, and generating a bald head image by inputting the face image to an artificial neural network that generates a head image in the face region; synthesizing the background image and the bald head image to generate a bald head background image; extracting a non-hair mask from which hair regions are removed by segmentation in the person image; and generating a bald person image by synthesizing the non-hair mask and the bald head background image.

The bald-headed person image generation method and other embodiments may include the following features.

According to an embodiment, the step of generating a face image by extracting a face region from the person image and generating a bald head image by inputting the face image to an artificial neural network that generates a head image in the face region includes: extracting a face image excluding the hair part based on the feature points of; and generating the bald head image by inputting the face image to an artificial neural network trained to generate a head portion from any input face image.

According to an embodiment, the artificial neural network may extract only a face region from a bald-headed person image with a background and a face area image generated by extracting only a face area and a background-free background generated by removing the background of the background-balanced person image with a segmentation model. When an arbitrary face region image is input to an image set consisting of a bald person image, it may be learned to generate a bald head image by generating a head portion in the face region.

According to an embodiment, the face image may be extracted based on a feature point corresponding to a face region among feature points predicted using the Dlib image processing tool.

Meanwhile, the present specification proposes a method for generating an image of a person with a bald head. The bald-headed person image generation method includes the steps of extracting a background mask from which a person object area is removed by performing segmentation on a person image, and filling in a portion from which the person object area is removed from the background mask with inpainting to generate a background image. ; generating a face image by extracting a face region from the person image, and generating a bald head image by inputting the face image to an artificial neural network that generates a head image in the face region; synthesizing the background image and the bald head image to generate a bald head background image; extracting a non-hair mask from which hair regions are removed by segmentation in the person image; extracting a bald head region corresponding to the hair region from the bald head background image based on the non-hair mask; and generating a bald person image by overlapping the bald head region on the character image.

The bald-headed person image generation method and other embodiments may include the following features.

According to an embodiment, the extracting of the bald head region corresponding to the hair region from the bald head background image based on the non-hair mask may include overlapping the non-hair mask on the bald head background image. Generating an overlapping image by doing; and extracting the bald hair region corresponding to the hair region from the overlapping image.

According to an embodiment, the step of generating a face image by extracting a face region from the person image and generating a bald head image by inputting the face image to an artificial neural network for generating a head image in the face region may include: Extracting a face image excluding the hair part based on the feature points of the face from; and generating the bald head image by inputting the face image to an artificial neural network trained to generate a head portion from any input face image.

According to an embodiment, the artificial neural network may extract only a face region from a bald-headed person image with a background and a face area image generated by extracting only a face area and a background-free background generated by removing the background of the background-balanced person image with a segmentation model. When an arbitrary face region image is input to an image set consisting of a bald person image, it may be learned to generate a bald head image by generating a head portion in the face region.

According to an embodiment, the face image may be extracted based on a feature point corresponding to a face region among feature points predicted using the Dlib image processing tool.

On the other hand, the present specification proposes an apparatus for generating an image of a person with a bald head. The bald-headed person image generating device may include a storage unit; and a control unit functionally connected to the storage unit, wherein the storage unit stores data generated by the control unit, and the control unit extracts a background mask from which a person object area is removed by subdividing a person image, , A background image is created by filling in the part where the person object area is removed from the background mask with inpainting, a face image is created by extracting a face area from the person image, and a head is created from the face image in the face area. A non-hair mask in which a bald head image is generated, a bald head background image is generated by synthesizing the background image and the bald head image, and segmentation is performed on the person image to remove the hair region. is extracted, and a first bald person image is generated by synthesizing the non-hair mask and the bald head background image, or from the bald head background image to the hair region based on the non-hair mask. A corresponding shaved head area may be extracted, and a second shaved head person image may be generated by overlapping the shaved head area on the person image.

The bare-headed person image generating device and other embodiments may include the following features.

According to an embodiment, the control unit extracts the face image excluding the hair part based on feature points corresponding to the face region among the feature points of the face predicted using the Dlib image processing tool in the person image, The bald head image may be generated by inputting the face image to an artificial neural network trained to generate a head portion from the face image.

According to an embodiment, the controller may overlap the non-hair mask on the bald head background image to generate an overlapped image, and extract the bald head region corresponding to the hair region from the overlapped image. there is.

According to an embodiment, the artificial neural network extracts only a face region from a bald-headed person image with a background and a face area image generated by extracting only a face region image and a bald-headed person without a background generated by removing the background of the background-balanced person image with a segmentation model. When an arbitrary face region image is input as an image set composed of images, it may be learned to generate a bald head image by generating a head portion in the face region.

On the other hand, this specification presents a computer program. The computer program may be combined with hardware and stored in a computer-readable recording medium to execute each step included in the above-described method for generating a bald person image.

Embodiments disclosed herein have an effect of generating a natural person image in which only hair is removed from a face image of a person with hair.

On the other hand, the effects obtainable in the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below. You will be able to.

The following drawings attached to this specification illustrate preferred embodiments of the present invention, and serve to further understand the technical idea of the present invention together with specific details for carrying out the invention, so the present invention is described in such drawings should not be construed as limited to
1 is a view conceptually illustrating a method of generating a bald person image by using a bald person image generating apparatus according to an exemplary embodiment.
2 shows an example of training data of a face segmentation deep learning model.
3 shows an example of a facial feature point extraction model.
4 shows an example of a method of generating a training data set for an artificial neural network that generates a head image of a bald head.
5 shows an example of bald face image data used for learning of an artificial neural network.
6 to 9 are flowcharts illustrating a method of generating an image of a person with a bald head using an apparatus for generating an image of a person with a bald head according to an exemplary embodiment.
10 is a configuration diagram of an apparatus for generating an image of a person with a bald head according to an exemplary embodiment.

The technology disclosed in this specification can be applied to an apparatus and method for generating a bald person image for a virtual hair styling experience. However, the technology disclosed in this specification is not limited thereto, and may be applied to all devices and methods to which the technical spirit of the technology may be applied.

It should be noted that technical terms used in this specification are only used to describe specific embodiments and are not intended to limit the spirit of the technology disclosed in this specification. In addition, technical terms used in this specification should be interpreted in terms commonly understood by those of ordinary skill in the field to which the technology disclosed in this specification belongs, unless specifically defined otherwise in this specification. It should not be interpreted in an overly comprehensive sense or in an excessively reduced sense. In addition, when the technical terms used in this specification are incorrect technical terms that do not accurately express the spirit of the technology disclosed in this specification, it is a technical term that can be correctly understood by those of ordinary skill in the field to which the technology disclosed in this specification belongs. should be replaced with In addition, general terms used in this specification should be interpreted as defined in advance or according to context, and should not be interpreted in an excessively reduced sense.

Terms including ordinal numbers such as first and second used herein may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention.

Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the accompanying drawings, but the same or similar components are assigned the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted.

In addition, in describing the technology disclosed in this specification, if it is determined that a detailed description of a related known technology may obscure the gist of the technology disclosed in this specification, the detailed description will be omitted. In addition, it should be noted that the accompanying drawings are only intended to facilitate understanding of the spirit of the technology disclosed in this specification, and should not be construed as limiting the spirit of the technology by the accompanying drawings.

Throughout the specification, a device or terminal includes a communication terminal or communication device capable of wired or wireless communication with a server or other device. The shape of the device or terminal may have various forms such as a mobile phone, a smart phone, a smart pad, a laptop computer, a desktop computer, a wearable device, a mirror-type display device, a smart mirror, and the like. The wearable device may be various such as a watch type terminal, a glasses type terminal, and an HMD. Also, the terminal is not limited to this form and may be implemented as various electronic devices.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

1 is a view conceptually illustrating a method of generating a bald person image by using a bald person image generating apparatus according to an exemplary embodiment.

2 shows an example of training data of a face segmentation deep learning model.

3 shows an example of a facial feature point extraction model.

1 to 3, the bald-headed person image generation device subdivides the person image 10 input into the device into parts, extracts a background mask obtained by removing an object area corresponding to the person in the person image, and extracts the extracted background. The background image 20 is created by filling the background from the mask with the object area removed through image inpainting. Here, a method of subdividing the input person image 10 into parts may be performed using a face segmentation deep learning model. The face segmentation deep learning model used in the embodiment may use a BiSeNet model pre-learned to produce a mask for each part when a face image is input. The BiSeNet model uses CelebAMask shown in FIG. 2 as training data.

In addition, the apparatus for generating a bald person image extracts a face area from the person image 10 to generate a face image 30, and then inputs the generated face image 30 to an artificial neural network that generates a head in the face area to have a bald head. A head image 40 is generated. Here, the method of generating the face image 30 by extracting the face region from the person image 10 predicts 68 facial keypoints using Dlib, an image processing tool as shown in FIG. , In order to remove unnecessary noise in making a bald head, the face image is created by cutting out the face area using feature points from 0 to 26 among the predicted feature points. On the other hand, the artificial neural network that generates the bald head image 40 uses NVIDIA's pix2pixHD model, called Image to Image Translation technology, as a deep learning model for generating a virtual head through the face area.

Next, the bald person image generating device generates a bald head image 50 with a background by synthesizing the background image 20 and the bald head image 40 .

In addition, the bald person image generating apparatus extracts a non-hair mask 60 from which the hair region 61 is removed from the previously segmented person image 10 .

Finally, the bald person image generating device synthesizes the non-hair mask 60 and the bald head image 50 having a background to generate a bald person image 70, or a non-hair mask ( 60), a region corresponding to the hair region 61 is extracted from the background image 50 of a bald head with a background, and the corresponding region extracted from the background image 50 of a bald head with a background is a person image ( 10) to create a bald person image 70.

Hereinafter, a learning method of an artificial neural network for generating an image of a bald head will be described with reference to the accompanying drawings.

4 shows an example of a method of generating a training data set for an artificial neural network that generates a head image of a bald head.

5 shows an example of bald face image data used for learning of an artificial neural network.

Referring to FIGS. 1, 4, and 5 , the apparatus for generating a bald person image 80 extracts only a face area based on the aforementioned facial feature points from a bald person image 80 with a background to generate a face area image 81, The background of the bald-headed person image 80 with a background is removed by the aforementioned face segmentation deep learning model to generate a bald-headed person image 82 without a background. Thereafter, an image set consisting of a face region image 81 and a bare-headed person image 82 without a background is included in training data. Next, when an arbitrary face region image 10 is input with the learning data generated in this way, the apparatus for generating a bald person image generates a head portion in the face region and trains the artificial neural network to generate a bald head image 40 . Here, the bald face image data shown in FIG. 5 is used as the bald person image 80 with a background.

6 to 9 are flowcharts illustrating a method of generating an image of a person with a bald head using an apparatus for generating an image of a person with a bald head according to an exemplary embodiment.

Referring to FIGS. 1 and 6 to 9 , the apparatus for generating a bald-headed person image according to an embodiment extracts a background mask from which a person object region is removed by subdividing a person image 10, and extracts the person object from the background mask. A background image 20 is created by filling the area from which the area has been removed with inpainting (S100). Here, the bald-headed person image generating device does not directly use the input person image, but uses a face crop algorithm to obtain a cropped face image obtained by cutting out only the face portion in a rectangle from the input person image as the person image 10. By using it, deep learning performance can be maximized.

Next, the shaved-headed person image generation device extracts a face region from the person image 10 to generate a face image 30, and inputs the face image 30 to an artificial neural network that generates a head image in the face region to generate a shaved head. A head image 40 is generated (S110). Here, the process of generating the bald head image 40 includes the process of extracting the face image 30 excluding the hair part based on facial feature points from the person image (S111) and the head image from any input face image. It may be performed by a method including a step of generating the bald head image 40 by inputting the face image 30 to an artificial neural network trained to generate the part (S112). Here, the face image 30 may be extracted based on feature points corresponding to the face region among feature points predicted using the Dlib image processing tool in the person image 10 . In addition, the artificial neural network consists of a facial area image generated by extracting only a face region from a bald-headed person image with a background and a bald-headed person image without a background generated by removing the background of the bald-headed person image with a background by a segmentation model. When an arbitrary face region image is input as an image set, it may be learned to generate a bald head image by generating a head portion in the face region.

Next, the bald person image generating device generates a bald head background image 50 by synthesizing the background image 20 and the bald head image 40 (S120). Here, the bald head background image 50 is an image generated by combining the background image 20 with the background of the bald head image 40 .

Next, the bare-headed person image generation apparatus extracts the mask 60 from which only the hair region 61 is removed from the person image 10 after segmentation is performed on the person image 10 (S130). A mask obtained by removing only the hair region from the person image 10 may be referred to as a non-hair mask.

Finally, the bald person image generating device generates a bald person image 70 by synthesizing the non-hair mask 60 and the bald head background image 50 (S140). Here, if a cropped face image obtained by cutting out only the face part from the original person image is used as the person image 10, the created bald person image 70 is superimposed on the original person image to make the person in the original person image bald. image can be created.

Meanwhile, referring to FIG. 7 , according to another embodiment, the apparatus for generating a bald person image generates a mask 60 obtained by removing only the hair region 61 from the person image 10 after segmentation is performed on the person image 10 . After extracting (S130), a bald head region corresponding to the hair region 61 is extracted from the bald head background image 50 based on the non-hair mask 60 (S150), and the extracted A bald head area may be overlapped on the person image 10 to generate a bald person image 70 (S160). Here, in the process of extracting the bald head region corresponding to the hair region (S150), the non-hair mask 60 is superimposed on the bald head background image 50 to create an overlapped image (S151), and It may be configured to include a process of extracting the bald head region corresponding to the hair region 61 from the overlapped image (S152).

In the foregoing description, the steps, procedures or actions may be further divided into additional steps, procedures or actions, or combined into fewer steps, procedures or actions, depending on the implementation of the invention. In addition, some steps, processes or operations may be omitted if necessary, or the order of steps or operations may be switched. In addition, each step or operation included in the above-described bare-headed person image generation method may be implemented as a computer program and stored in a computer-readable recording medium, and each step, process, or operation may be executed by a computer device.

Hereinafter, with reference to the accompanying drawings, a bare-headed person image generating device will be described.

10 is a configuration diagram of an apparatus for generating an image of a person with a bald head according to an exemplary embodiment.

Referring to FIGS. 1 and 10 , the bald-headed person image generating apparatus 100 may include a communication unit 110 , a storage unit 120 , and a control unit 130 . The illustrated components are not essential, so the bare-headed person image generating device 100 having more or fewer components may be implemented. These components may be implemented in hardware or software, or through a combination of hardware and software.

The communication unit 110 may transmit/receive data with an external device through a network.

The storage unit 120 may store data generated by the controller 130 or data received through the communication unit 110 .

The control unit 130 is functionally connected to the communication unit 110 and the storage unit 120 .

In addition, the controller 130 performs subdivision on the person image 10 to extract a background mask from which the person object area is removed, and fills in the background mask from which the person object area is removed with inpainting to fill the background image 20 with inpainting. ) can be created.

In addition, the controller 130 generates a face image 30 by extracting a face region from the person image 10, inputs the face image 30 to an artificial neural network that generates a head in the face region, and generates a bald head image 40. ) can be created.

In addition, the controller 130 may generate a bald head background image 50 by synthesizing the background image 20 and the bald head image 40 .

In addition, the controller 130 may perform segmentation on the person image 10 to extract a non-hair mask 60 from which the hair region 61 is removed.

In addition, the controller 130 generates a bald person image 70 by synthesizing the non-hair mask 60 and the bald head background image 50, or based on the non-hair mask 60 A bald person image 70 may be generated by extracting a bald head region corresponding to the hair region 61 from the bald head background image 50 and overlapping the extracted bald head region on the person image 10 . Here, the controller 130 overlaps the non-hair mask 60 on the bald head background image 50 to generate an overlapped image, and extracts a bald head region corresponding to the hair region 61 from the overlapped image. can

In addition, the controller 130 extracts the face image 30 excluding the hair part based on the feature points corresponding to the face area among the feature points of the face predicted using the Dlib image processing tool in the person image 10, and extracts the face image 30 excluding the hair part. The bald head image 40 may be generated by inputting the face image 30 to an artificial neural network trained to generate a head portion from an arbitrary face image. Here, the artificial neural network consists of a face area image generated by extracting only the face region from a bald-headed person image with a background and a bald-headed person image without a background generated by removing the background of the bald-headed person image with the background by a segmentation model. When an arbitrary face region image is input as an image set, the control unit 130 learns to generate a bald head image by generating a head portion in the face region.

A network disclosed herein may include, for example, a wireless network, a wired network, a public network such as the Internet, a private network, a Global System for Mobile communication network (GSM) network, a general packet wireless network (General Packet Radio Network (GPRN), Local Area Network (LAN), Wide Area Network (WAN), Metropolitan Area Network (MAN), Cellular Network, Public Switched Telephone Network PSTN), Personal Area Network, Bluetooth, Wi-Fi Direct, Near Field communication, Ultra-Wide band, a combination thereof, or any It may be other networks, but is not limited thereto.

As used herein, the term “unit” (eg, a controller) may refer to a unit including one or a combination of two or more of, for example, hardware, software, or firmware. “Unit” may be used interchangeably with terms such as, for example, unit, logic, logical block, component, or circuit. A "unit" may be a minimum unit of an integrally constituted part or a part thereof. A “unit” may be a minimal unit or part thereof that performs one or more functions. A “unit” may be implemented mechanically or electronically. For example, a "part" is an application-specific integrated circuit (ASIC) chip, field-programmable gate arrays (FPGAs), or programmable-logic device that performs certain operations, known or developed in the future. may contain at least one.

At least some of the devices (eg, modules or functions thereof) or methods (eg, operations) according to various embodiments may be stored on computer-readable storage media in the form of, for example, program modules. It can be implemented as a command stored in . When the command is executed by a processor, the one or more processors may perform a function corresponding to the command. The computer-readable medium includes all types of recording devices in which data that can be read by a computer system is stored. Computer-readable storage media / computer-readable recording media include hard disks, floppy disks, magnetic media (eg magnetic tape), optical media (eg CD-ROM (compact) disc read only memory), digital versatile disc (DVD), magneto-optical media (e.g. floptical disk), hardware devices (e.g. read only memory (ROM), random access memory), or flash memory, etc.), and may also include those implemented in the form of carrier waves (eg, transmission through the Internet) In addition, program commands include machine words such as those created by a compiler. code as well as high-level language codes that can be executed by a computer using an interpreter, etc. The above-described hardware device may be configured to operate as one or more software modules to perform operations of various embodiments, and vice versa. Same.

A module or program module according to various embodiments may include at least one or more of the aforementioned elements, some may be omitted, or additional elements may be further included. Operations performed by modules, program modules, or other components according to various embodiments may be executed in a sequential, parallel, repetitive, or heuristic manner. Also, some actions may be performed in a different order, omitted, or other actions may be added.

As used herein, the term "a" is defined as one or more than one. Also, the use of introductory phrases such as “at least one” and “one or more” in a claim means that the same claim includes introductory phrases such as “at least one” and “one or more” and ambiguous phrases such as “an”. If any, be construed to mean that the introduction of another claim element by the ambiguous phrase "an" limits any particular claim containing the so-introduced claim element to an invention containing only one such element. It shouldn't be.

In this document, expressions such as "A or B" or "at least one of A and/or B" may include all possible combinations of the items listed together.

Unless otherwise specified, terms such as "first" and "second" are used to arbitrarily distinguish the elements they describe. Thus, these terms are not necessarily intended to indicate a temporal or other order of priority of such elements, and the mere fact that certain measures are recited in mutually different claims does not indicate that a combination of these measures cannot be used to advantage. . Accordingly, these terms are not necessarily intended to indicate a temporal or other priority of such elements. The mere fact that certain measures are cited in different claims does not indicate that a combination of these measures cannot be useful.

Arrangements of components to achieve the same function are effectively "related" so that the desired function is achieved. Thus, any two components that are combined to achieve a particular functionality may be considered "related" to each other such that the desired function is achieved, regardless of structure or intervening components. Similarly, two components so associated can be considered "operably connected" or "operably coupled" to each other to achieve a desired function.

Further, those skilled in the art will recognize that the boundaries between the functionality of the foregoing operations are exemplary only. A plurality of actions may be combined into a single action, a single action may be distributed into additional actions, and the actions may be executed at least partially overlapping in time. Also, alternative embodiments may include multiple instances of a particular operation, and the order of operations may be changed in various other embodiments. However, other modifications, variations and alternatives are also possible. Accordingly, the detailed description and drawings are to be regarded in an illustrative rather than a limiting sense.

The phrase “may be X” indicates that condition X can be met. This phrase also indicates that condition X may not be met. For example, a reference to a system that contains a specific component should also include a scenario where the system does not contain that specific component. For example, a reference to a method that includes a specific action must also include scenarios in which the method does not include that specific component. But to take another example, a reference to a system configured to perform a specific action should also include a scenario in which the system is not configured to perform the specific action.

The terms "comprising", "having", "consisting of", "consisting of" and "consisting essentially of" are used interchangeably. For example, any method may include at least the operations included in the drawings and/or specifications, and may include only the operations included in the drawings and/or specifications. Alternatively, the word "comprising" does not exclude the presence of the recited elements or acts in a claim.

It should be understood that the boundaries between logical blocks are merely illustrative, and that alternative embodiments may merge logical blocks or circuit elements or impose alternative decompositions of functionality on various logical blocks or circuit elements. will recognize that there is Accordingly, it should be understood that the architecture shown herein is merely illustrative, and in fact many other architectures that achieve the same functionality may be implemented.

Also, for example, in one embodiment, the illustrated examples may be implemented on a single integrated circuit or as circuitry located within the same device. Alternatively, the above examples may be implemented as any number of discrete integrated circuits or discrete devices interconnected with each other in any suitable manner, and other alterations, modifications, variations and alternatives are also possible. Accordingly, the specification and drawings are to be regarded in an illustrative rather than restrictive sense.

Also, for example, any of the foregoing examples or portions thereof may be implemented as a software or code representation of a physical circuit or a logical representation translatable to a physical circuit, such as any suitable type of hardware description language.

In addition, the present invention is not limited to physical devices or units implemented with non-programmable hardware, but generally includes mainframes, mini computers, servers, workstations, personal computers, notepads, referred to herein as 'computer systems'. Programmable devices capable of performing desired device functions by operating in accordance with appropriate program code, such as personal digital assistants (PDAs), electronic games, automobiles and other embedded systems, mobile phones and various other wireless devices. Or it can be applied to units as well.

A system, apparatus or device referred to in this specification includes at least one hardware component.

Connections as described herein may be any type of connection suitable for transmitting a signal from or to a respective node, unit or device via an intermediate device, for example. Thus, unless implicitly or otherwise stated, a connection may be, for example, a direct connection or an indirect connection. A connection may be described or described with reference to a single connection, multiple connections, unidirectional connections, or bidirectional connections. However, different embodiments may change the implementation of connectivity. For example, separate unidirectional connections can be used instead of bidirectional connections, and vice versa. Also, multiple connections may be replaced with a single connection that transmits a plurality of signals sequentially or in a time multiplexed manner. Likewise, a single connection carrying multiple signals can be broken into multiple connections carrying subsets of these signals. Therefore, many options exist for transmitting signals.

In the above, a preferred embodiment of the technology of this specification has been described with reference to the accompanying drawings. Here, terms or words used in this specification and claims should not be construed as being limited to ordinary or dictionary meanings, but should be interpreted as meanings and concepts consistent with the technical spirit of the present invention. The scope of the present invention is not limited to the embodiments disclosed herein, and the present invention can be modified, changed, or improved in various forms within the spirit of the present invention and the scope described in the claims.

10: Portrait image
20: background image
30: face image
40: bald head image
50: bald head background image
60: non-hair mask
61: hair area
70: Bald head portrait image
100: bald-headed person image generating device
110: communication department
120: storage unit
130: control unit

Claims (14)

In the method of generating a bald person image by a bald person image generating device,
segmenting the person image to extract a background mask from which the person object area is removed, and filling in the area from which the person object area is removed from the background mask with inpainting to generate a background image;
A face image without a background is generated by extracting a face region from the person image, and a bald head image without a background is generated by inputting the face image without a background to an artificial neural network that generates a head image in the face region. A face image without a background except for the hair part is extracted based on the feature points of the face, and a face region image without a background and a background of the face region image without a background and a background of the face image with the background extracted only the face region from the bald-headed person image with the background segmentation model When an arbitrary face area image is input to an image set consisting of a bare-headed person image without a background removed by , the artificial neural network that has been trained to create a bald-headed head image without a background by generating a head portion in the face area without the background generating a bald head image without the background by inputting a face image;
generating a bald head image with a background by synthesizing the background image and the bald head image without a background;
extracting a non-hair mask from which hair regions are removed by segmentation in the person image; and
generating a bald person image by synthesizing the bald head image with the non-hair mask and the background;
How to create an image of a bald-headed person. delete delete The method of claim 1, wherein the face image without a background,
Among the feature points predicted using the Dlib image processing tool, extracted based on the feature point corresponding to the face area
A method for generating a bald person image, characterized in that. In the method of generating a bald person image by a bald person image generating device,
segmenting the person image to extract a background mask from which the person object area is removed, and filling in the area from which the person object area is removed from the background mask with inpainting to generate a background image;
A face image without a background is generated by extracting a face region from the person image, and a bald head image without a background is generated by inputting the face image without a background to an artificial neural network that generates a head image in the face region. A face image without a background except for the hair part is extracted based on the feature points of the face, and a face region image without a background and a background of the face region image without a background and a background of the face image with the background extracted only the face region from the bald-headed person image with the background segmentation model When an arbitrary face area image is input to an image set consisting of a bare-headed person image without a background removed by , the artificial neural network that has been trained to create a bald-headed head image without a background by generating a head portion in the face area without the background generating a bald head image without the background by inputting a face image;
generating a bald head image with a background by synthesizing the background image and the bald head image without a background;
extracting a non-hair mask from which hair regions are removed by segmentation in the person image;
extracting a bald head region corresponding to the hair region from the bald head image with the background based on the non-hair mask; and
Generating a bald person image by overlapping the bald head region on the character image;
How to create an image of a bald-headed person. According to claim 5,
The step of extracting a bald head region corresponding to the hair region from the bald head image with the background based on the non-hair mask,
generating an overlapping image by overlapping the non-hair mask on the background image of the bald head; and
Extracting the bald head region corresponding to the hair region from the overlapping image;
A method for generating a bald person image, characterized in that. delete delete The method of claim 5, wherein the face image without the background,
Among the feature points predicted using the Dlib image processing tool, extracted based on the feature point corresponding to the face area
A method for generating a bald person image, characterized in that. storage unit; and
Including; a control unit functionally connected to the storage unit;
The storage unit stores data generated by the controller,
The control unit,
segmentation is performed on the person image to extract a background mask from which the person object area is removed, and a background image is created by filling in the part from which the person object area is removed from the background mask with inpainting;
A face image without a background is generated by extracting a face region from the person image, and a bald head image without a background is generated by inputting the face image without a background to an artificial neural network that generates a head image in the face region. A face image without a background except for the hair part is extracted based on the feature points of the face, and a face region image without a background and a background of the face region image without a background and a background of the face image with the background extracted only the face region from the bald-headed person image with the background segmentation model When an arbitrary face area image is input to an image set consisting of a bare-headed person image without a background removed by , the artificial neural network that has been trained to create a bald-headed head image without a background by generating a head portion in the face area without the background Inputting a face image to generate a bald head image without the background;
synthesizing the background image and the bald head image without a background to generate a bald head image with a background;
Segmentation is performed on the person image to extract a non-hair mask from which the hair region is removed;
A first bald person image is generated by synthesizing the non-hair mask and the bald head image with the background, or the hair region is generated from the bald head image with the background based on the non-hair mask. Extracting a corresponding bald head area and overlapping the bald head area on the person image to generate a second bald person image
An apparatus for generating images of bald-headed people. delete 11. The method of claim 10, wherein the control unit,
Overlaying the non-hair mask on the bald head background image to generate an overlapping image, and extracting the bald head region corresponding to the hair region from the overlapping image
An apparatus for generating a bald person image, characterized in that. delete combined with hardware
A computer program stored in a computer-readable recording medium to execute each step included in the method of generating a bald-headed person image according to any one of claims 1, 4 to 6, and 9.

Images