WO2020079235A1

WO2020079235A1 - Method and apparatus for accessing clothing

Info

Publication number: WO2020079235A1
Application number: PCT/EP2019/078396
Authority: WO
Inventors: Sergio FREGONI; Filippo VEZZALI; Tommaso ALESSI; Guglielmo UBALDI; Gabriele TAZZARI; Federico Marchetti
Original assignee: Yoox Net-A-Porter Group S.P.A.
Priority date: 2018-10-19
Filing date: 2019-10-18
Publication date: 2020-04-23

Abstract

The present invention relates to a computer-implemented method of providing access to a stock of clothing from mood board data, including training an AI engine based on the mood board data to generate a clothing; accessing a database of a stock of clothing items available through a managed supply chain; selecting from the database a plurality of clothing items of different types for combination together as one or more outfits; and causing to be displayed image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

Description

Method and Apparatus for Accessing Clothing

[0001] This invention relates to a method and system of apparatus for providing access to a stock of clothing.

BACKGROUND

[0002] Conventionally, when a person requires clothing, such as for an occasion, the person may consider the sort of clothing they want to wear for the occasion and seek out suitable clothing, for example for purchase. There may be hundreds if not thousands and sometimes tens of thousands of different items of clothing from which to choose.

[0003] Purchase of the clothing can be in a physical retail shop, or using an online e- commerce shop. Sometimes, the person may enlist the help of another person, for example a personal shopper or a friend, to help them identify suitable clothing items for the occasion, including help seeking out suitable clothing for purchase. Typically, the person will then consider each of the items of suitable clothing, before deciding whether to purchase one or more of the items of clothing.

[0004] In order to identify clothing suitable for the occasion, many different unsuitable clothing items must be manually considered and discarded by the person or the another person (or both). This process is labour-intensive, and also requires knowledge of all of the different items of clothing available in a stock of clothing.

[0005] In a physical shop, the person can try-on outfits of clothing themselves and form an assessment of whether the outfit is suitable for their needs. Sometimes, the person can be assisted by a sales assistant, a friend or a personal shopper. When satisfied, the user can purchase the selected items of clothing to form the outfit.

[0006] In an online e-commerce shop, the user typically has access to pictures of a large number of items of clothing. The user can reduce the number of items of clothing using filters. Nevertheless, it is typical for a user to order a number of different items of clothing to try at home, and then return any items of clothing that are unsuitable. This leads to inefficiencies in the supply chain, with items of clothing travelling to potential customers, then having to be reprocessed when they are returned.

[0007] The present disclosure seeks to provide at least an alternative to a method of providing access to a stock of clothing of the prior art, including access to one or more outfits from the stock of clothing.

BRIEF SUMMARY OF THE DISCLOSURE

[0008] In accordance with an aspect of the present disclosure there is provided a computer-implemented method of providing access to a stock of clothing. The method comprises: receiving mood board data indicative of a design theme of a mood board; training an Artificial Intelligence, Al, engine based on the mood board data to generate a clothing item selection machine characteristic of the design theme for selecting clothing items deemed to correspond to the design theme; accessing a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing; selecting from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme; locating clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and causing to be displayed on a display of a user terminal image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

[0009] Thus, the stock of clothing can be broken down and presented to a user in a manageable subset, based on a design theme embodied in a mood board. This allows access to the stock of clothing by presenting the user with suitable items from the stock of clothing. Furthermore, by presenting the user with clothing items for combination together as one or more outfits from the stock of clothing, instead of individual clothing items in isolation, the user is presented with clothing items that can be put together to form a complete outfit and also correspond to the design theme.

[0010] As used herein, the term“clothing items” will be understood to include any items to be worn or presented as part of an outfit of a user, including items of apparel. Clothing items are not limited to items of apparel to be worn on the torso or legs, but also include jewellery, headwear, shoes, bags and other accessories that can be considered part of an outfit. With reference to clothing items, the term“type” will be understood to mean a particular category of clothing item. For example, the type could be an item worn on a particular part of the body, of a particular style or of a particular material. Different types of clothing include clothing on different parts of the body, for example, hats, scarves, jewellery, shoes, bags and dresses. Different types of clothing may also include different types of clothing on a particular part of the body. For example, for the torso different types of clothing may include jumpers, t-shirts, vest tops, shirts, and cardigans.

[0011] The predetermined pose may be substantially any pose, for example a standing pose, where the person is standing up straight. In the predetermined pose, both arms may be resting substantially straight at each side of the person. It will be understood that other poses may also be used.

[0012] At least two of the plurality of selected clothing items may be rendered as being worn by the avatar.

[0013] The mood board data may comprise at least one image indicative of the design theme. The at least one image may comprise an image of clothing. Thus, the images of clothing that are considered to correspond to the design theme can be part of the mood board. The Al engine can be trained to select outfits containing similar clothing or similar styles of clothing to those forming part of the mood board. The at least one image may comprise an image not of clothing. Thus, items other than clothing may be used to define the design theme.

[0014] The method may further comprise generating, using the clothing item selection machine, one or more outfits that correspond to the design theme as a combination of a plurality of selected clothing items of different types. The outfits may be for displaying on the display of a user terminal. Thus, the outfits themselves can also be generated. Thus user can be presented with the generated outfits corresponding to the design theme.

[0015] The method may further comprise causing to be displayed on the display of the user terminal image data representative of at least one further outfit combining a plurality of the selected clothing items, wherein at least one clothing item associated with each further outfit is rendered as being worn by the avatar. Thus, a plurality of different outfits can be presented to the user, and each outfit can be displayed on the avatar. This allows a user to more easily compare the various outfits.

[0016] The method may further comprise causing to be displayed sequentially on the display of the user terminal image data representative of a series of outfits of selected clothing. In this way, the user can be presented with a choice of outfits, all presented together in a series.

[0017] The method may further comprise causing to be displayed on a display of a user terminal in juxtaposition to the avatar and the outfit, other selected clothing items. Thus, not all selected clothing items need be displayed on the avatar. Furthermore. Clothing items other than those forming the outfit may be displayed to the user. The method may further comprise causing to be displayed on a display of a user terminal, in response to user input on the user terminal in relation to choosing one of the displayed other selected clothing items, image data representative of the avatar and of an adapted outfit including the user-chosen clothing item. Thus, the clothing items displayed in juxtaposition to the avatar and the outfit may be displayed as part of the outfit when selected. This allows the user to change the clothing items displayed on the avatar and as part of the outfit. [0018] The avatar may be substantially photo-realistic, such that the representation of the at least one item of clothing on the avatar is suitable to illustrate how the item of clothing would appear on a person. Thus, the avatar has proportions that correspond to those which would be possible on a real person. In other words, the avatar does generally not have unreal and/or caricatured proportions. Thus, the user can see how the outfit would look on an avatar indicative of the proportions and appearance of a person.

[0019] The avatar may be representative of a real person. Thus, the user can see how the outfit would appear on a real person. The real person may be the user of the user terminal. In this way, the user can see how the outfit would appear on themselves.

[0020] Rendering the at least one of the plurality of clothing items on the avatar may comprise combining the representations of the plurality of clothing items with a

representation of the avatar, such that the avatar appears to be wearing the at least one clothing item. Thus, the clothing need not be combined on the avatar prior to the representation of the clothing items being acquired. In this way, representations of the clothing items can be acquired separately to the choice of avatar on which to represent the plurality of clothing items.

[0021] Causing image data to be displayed on a display of a user terminal may comprise combining the image data representative of the avatar with the image data representative and geometrically transforming the image of the or each at least one clothing item of the plurality of clothing items to appear fitted to the representation of the person, and rendering the combined image data. It will be understood that the term geometrically transforming means any modification of at least the size and shape of any region of the image of the clothing item.

[0022] The method may further comprise causing to be displayed on a display of a user terminal in juxtaposition to the avatar and the outfit an indication of the design theme with which the outfit is associated. Thus, the user can be informed about the design theme with which the outfit is associated.

[0023] At least one of the plurality of clothing items may be displayed on the electronic display of the user terminal off the avatar, adjacent to the avatar. Thus, not all items of the outfit need be displayed on the avatar. For example, for small items of clothing, or where complex image processing would be needed to present the clothing item on the avatar, the item of clothing may be better presented to the user by being presented off the avatar. In the case of small items of clothing, this is because the item of clothing can be presented at a larger size.

[0024] The at least one of the plurality of clothing items displayed adjacent to the avatar may comprise at least one of: footwear, a bag, jewellery and a wallet and a purse. [0025] The method may comprise creating mood board data indicative of the design theme. In one example, a merchandiser may choose a plurality of images that they consider correspond to the design theme. These images may form the mood board data.

[0026] The mood board data may be created based on one or more data inputs received at a device other than the user terminal. Thus, the mood board may be curated by a person other than the user, for example, a professional clothing stylist. In this way, users can benefit from the design skills of a highly skilled clothing style specialist, without requiring the clothing stylist to be familiar with the details of an individual user.

[0027] The clothing item selection machine may identify the clothing items further based on a user selection input in relation to one or more of the clothing items in the stock of clothing. Thus, a merchandiser or other user can supervise an initial selection from the Al engine used as part of the clothing item selection machine. In this way, any clothing items that do not correspond to the merchandiser’s vision of the design theme can be removed from the selected clothing items. Furthermore, even when the clothing item selection machine identifies a large number of clothing items which can be combined together to form outfits which correspond to the design theme, the merchandiser may select only a subset of the clothing items as the final output from the clothing item selection machine.

[0028] The user selection input may be used to further train the Al engine. Thus, the merchandiser (or the user)’s input can be used to improve the accuracy of the Al engine. The performance of the Al engine can be improved over time.

[0029] The method may further comprise receiving user characteristic data indicative of one or more user characteristics including user size characteristics. Selecting from the database a plurality of clothing items may further be based on the user characteristic data. In this way, the exact selection of clothing items may be tailored to the specific

characteristics of the user. The user characteristics may include size, for example at least one of waist, chest and height. The user characteristics may include other filters, such as price range.

[0030] Selecting from the database a plurality of clothing items may further be based on inventory of deliverable stock of the clothing items matching the user characteristic data. Thus, whether the clothing items can be delivered to the user may be taken into account; only clothing items which can be delivered to the user may be selected for presentation to the user.

[0031] The further outfit may comprise at least one retained clothing item also present in the first outfit, and a further clothing item different to any of the clothing items associated with the first outfit, and wherein the representation of each of the plurality of clothing items associated with the second outfit is caused to be displayed on the electronic display in response to a change request received from the user, wherein the change request is indicative of a request from the user to change a clothing item in the first outfit. Thus, the user can change the items of clothing to be presented on the electronic display if, for example, one of the clothing items in the plurality of clothing items is not to the taste of the user.

[0032] The method may further comprise causing to be displayed on the display, image data representative of a further avatar of a further person in the predetermined pose instead of the avatar in response to an avatar change request received at the user terminal, wherein the at least one clothing item is rendered as being worn by the further avatar. Thus, the user of the user terminal can view the outfit on an avatar more suited to their needs, for example more similar to themselves.

[0033] The method may further comprise sharing data indicative of the image data representative of the avatar of the person in the predetermined pose, having rendered thereon the at least one clothing item to a further device in response to a sharing request received at the user terminal. The further device may be a further user terminal of a further person. The shared data may comprise the image data representative of the outfit combining the plurality of the selected clothing items. Thus, the user can share the outfit made up of the selected clothing items to a third party, for example a friend. The data may be shared via a social media application, for example one or more of Facebook (RTM), Instagram (RTM) and Snapchat (RTM).

[0034] The database of the stock of clothing may comprise at least 1000 different clothing items. Thus, the database of the stock of clothing would otherwise be relatively unwieldy and difficult for the user to access.

[0035] Viewed from another aspect, the present disclosure provides a system of apparatus for providing access to a stock of clothing. The system comprises: a server configured to: receive mood board data indicative of a design theme of a mood board; train an Artificial Intelligence, Al, engine based on the mood board data to generate a clothing item selection machine characteristic of the design theme for selecting clothing items deemed to correspond to the design theme; access a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing; select from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme; locate clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and cause display data to be transmitted to a user terminal; and a user terminal configured to: display, on a display of a user terminal and in response receipt from the server of display data, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar. Thus, the method described hereinbefore can be carried out on a server and user terminal. It will be understood that any of the further features and steps of the method described herein can also be carried out on either one or both of the server or user terminal or any further device as described herein.

[0036] The disclosure extends to a server configured as recited with reference to the system of apparatus hereinbefore.

[0037] The disclosure extends to a user terminal configured as recited with reference to the system of apparatus hereinbefore.

[0038] In accordance with a further aspect of the present disclosure, there is provided a computer-implemented method of providing access to a stock of clothing. The method comprises: accessing a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing; selecting from the database a plurality of clothing items of different types by using a clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits; locating clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and causing to be displayed on a display of a user terminal image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Embodiments of the invention are further described hereinafter with reference to the accompanying drawings, in which:

Figure 1 is an example flowchart of a method of providing access to a stock of clothing.

Figure 2A is a schematic diagram illustrating an apparatus according to an example.

Figure 2B is a schematic diagram illustrating an apparatus according to another example.

Figures 3A and 3B illustrate examples of mood boards and clothing items.

Figures 4A and 4B provide schematic examples of a display of a user terminal. Figures 5A and 5B provide illustrative examples of a display of a user terminal.

Figures 6A and 6B provide further schematic examples of a display of a user terminal.

Figures 7A and 7B provide further illustrative examples of a display of a user terminal.

DETAILED DESCRIPTION

[0040] The present disclosure describes a computer-implemented method of providing access to a stock of clothing. The method comprises: receiving mood board data indicative of a design theme of a mood board; training an Artificial Intelligence, Al, engine based on the mood board data to generate a clothing item selection machine, characteristic of the design theme, for selecting clothing items deemed to correspond to the design theme; accessing a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing; selecting from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme; locating clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and causing to be displayed on a display of a user terminal image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

[0041] Figure 1 is an example flowchart of a method 100 of providing access to a stock of clothing. In the method 100, mood board data is received 102. The mood board data is indicative of a design theme of a mood board. The mood board data will be described later with reference to figures 3A and 3B. After receiving 102 the mood board data, an Al engine is trained 104 based on the mood board data. The Al engine is trained 104 to generate a clothing item selection machine, characteristic of the design theme. The clothing item selection machine is for selecting clothing items deemed to correspond to the design theme. A database of a stock of clothing items available through a managed supply chain is accessed 106 as part of the method 100. The database of the stock of clothing items comprises clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing. After accessing 106 the database of the stock of clothing items, a plurality of clothing items of different types is selected 108 from the database by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme. Clothing image data representative of the selected clothing items is then located 110. The clothing image data for at least some of the types of clothing is representative of the clothing items as they would appear when worn by a wearer in a predetermined pose. Next, the image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items is caused to be displayed 112 on a display of a user terminal. At least one clothing item is rendered as being worn by the avatar. This method will be described in further detail later with reference to figures 3A and 3B.

[0042] Figure 2A is a schematic diagram illustrating an apparatus 200 according to an example. The apparatus 200 includes a server 202 and a user terminal 204. The apparatus 200 may perform the method 100 of figure 1. The method steps of figure 1 may be performed in the server 202 and/or the user terminal 204 in substantially any combination. The user terminal 204 may include a display 206 or the display 206 may be separate from the user terminal 204 and connected to the user terminal 204 via a wireless connection or a wired connection. In this example, the server 202 is configured to receive mood board data indicative of a design theme of a mood board; train an Al engine based on the mood board data to generate a clothing item selection machine, characteristic of the design theme, for selecting clothing items deemed to correspond to the design theme; access a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more

characteristics of each of the clothing items in the stock of clothing; select from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme; locate clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and cause display data to be transmitted to a user terminal 204. In this example, the user terminal 204 is configured to display, on the display 206 of the user terminal 204 and in response receipt from the server 202 of display data, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

[0043] The display 206 may be an electronic display for displaying image data thereon for viewing by a user of the user terminal 204.

[0044] In an example, user terminal 204 may be able to receive user inputs. For example, user terminal 204 may include one or more input buttons. In some examples, the display 206 associated with user terminal 204 may be a touch screen display and user terminal 204 may receive user inputs via the touch screen display. This apparatus will be described in further detail later with reference to figures 3A and 3B.

[0045] Figure 2B is a schematic diagram illustrating an apparatus 250 according to another example. Apparatus 250 comprises server 252 and user terminal 254. The server 252 is an example of the server 202 of apparatus 200 and the user terminal 254 is an example of the user terminal 204 of apparatus 200. In an example, user terminal 254 comprises display 256, processor 266 and memory 268. It will be understood that the user terminal 204 can comprise other components not shown, such as more processors and a communication module. The communication module is typically for data communication with the server 202. Display 256 is an example of display 206 of apparatus 200. The memory 258 may comprise instructions, which when executed by the one or more processors 266, cause the user terminal 204 to perform steps of the herein-described methods. Memory 268 may store image data and/or video data for displaying on the display 256. In an example, memory 268 stores, either temporarily or permanently, the data received from the server 252, for example, clothing image data. Memory 268 may comprise instructions for the display 256 or to enable the user terminal 254 to

communicate with the server 252. In an example, user terminal 254 is a handheld device, for example, a mobile phone or tablet. The memory 268 may be a computer-readable non- transitory storage medium. In some examples, the memory 268 may comprise a plurality of separate memories 268, the image data and/or video data for displaying on the display 256 and the instructions for causing the one or more processors 266 to carry out the herein-before described methods, may together be stored on the plurality of memories 268.

[0046] In an example, server 252 comprises memory 258, an Al engine 262 and a clothing item selection machine 264. Server 252 may also include a processor (not shown). Memory 258 may include a database of a stock of clothing items 260. The server 252 may be a cloud server and may be located in the“cloud”. For example, the server may be located anywhere where it can still be communicative with the user terminal 254.

Moreover, the server 252 may be part of a shared computing system. The memory 258, Al engine 262 and clothing item selection machine 264 may be located separately, for example, in different servers 252. The Al engine 262 may be trained to produce an algorithm which it then supplies to the clothing item selection machine 264 to assist the clothing item selection machine 264 with clothing item selection. In another example, the Al engine 262 and clothing item selection machine 264 may form one module. For example, the clothing item selection machine 264 may comprise the Al engine 262. [0047] Figures 3A and 3B illustrate examples 300, 350 of mood boards 302, 352 and clothing items 408 to 426 and 458 to 476. As used herein, mood board data is data that at least contributes to the formation of a mood board 302, 352. For example, mood board data is data indicative of the mood board 302, 352. For example, mood board data may include a plurality of images, each part of the mood board. For example, the images may contain models, still life, patterns or places. For example, the images may relate to products from a seller’s catalogue. Mood board data may be representative of at least a part of the mood board. In an example, mood board data may be the mood board as it appears in memory. Thus, any further references to a mood board refer to the mood board and/or the mood board data.

[0048] A mood board may be designed by a designer based on a design theme. Thus, the mood board may be based on the designer’s understanding of a design theme but also may be based on the designer’s preferences, the trend when the mood board is created and the target audience of such a mood board. Typically, the mood board is an artistic piece and is not created by a computer. Generation of a mood board, including images representative of a design theme is carried out by the designer. Additionally or

alternatively, a computer may be used to assemble mood boards. For example, the mood board may be at least partially generated by a computer.

[0049] A user solely relying on a computer to select their clothing could be presented with either a mostly random mixture of clothes that do not match a theme or a set of clothes of the same type or that would all be matching in some way. This does not present the user with a variety of clothing for a particular theme, as the user so desires, and is thus dissatisfactory.

[0050] It is known in the clothing and fashion industry that there are a great wealth of types of clothing options, and that many of these clothing options may not be combinable in a manner that would satisfy certain fashion trends or dress codes. For example, tracksuit bottoms or joggers are a type of pant/trouser, but most people would consider them not to be an acceptable type of pant for a formal occasion. Such rules may extend beyond types of clothing, as there may be colours, materials or patterns that are widely considered to“clash” and thus not to be combinable in an outfit. This imposes a large number of constraints on clothing combinability. Moreover, some of these constraints are subjective, or may vary based on factors including a particular person’s occupation/career, location, or demographics (e.g., age, ethnicity, etc.). Those skilled in the art of machine learning and artificial intelligence will recognize that when there are a large number of constraints, this imposes a significant impediment to training an automated system to successfully carry out a task. [0051] Attempts have previously been made to construct automated systems to help with clothing selection. Given the large number of constraints mentioned above, these prior approaches have focused on having knowledgeable users, such as designers, manually input rules regarding combinability of clothing items. Such rules may specify what types of clothing may be worn in different parts of the body under different dress codes. For example, in such a system, a rule may specify that for men’s formal businesswear, a torso region of the body could accommodate both a dress shirt and a suit jacket. The rules may also specify alternatives for clothing items within a particular dress code, such as that one type of shoe is an alternative to another type of shoe, but that either type may be worn in accordance with a certain dress code. Once the manual rules regarding potential types of clothing are input, a large number of manually-encoded images of clothing are input, including in combinations. The manual encoding indicates further information about clothing that may be combined, such as colours, patterns, etc. In some cases, further explicit manual input may be provided to train the system for factors such as ethnicity/skin tone.

[0052] Such systems may operate to automatically source clothing options that may be combinable in an outfit in accordance with the manually-input rules and other training, and some such systems may operate acceptably once trained. However, there are

disadvantages to these systems. Among the disadvantages is that they require great manual effort to explicitly define rules relating to the types of clothing that may be worn, e.g., within a certain dress code and relating to combinations of those clothing items. In addition, once the rules are defined, a great deal of manually-encoded examples need be further input, to additionally train the system to operate well. As discussed above, these prior systems enabled input of this information to try to account for the great number and variety of constraints on clothing selection, and operate well in spite of these constraints— and may be behind why some such systems may work well once trained. However, such a manual effort also heavily limits the usability of such systems, as the manual effort imposes a great burden on creation of new dress codes, i.e., that detailed rules and manually-encoded examples must be specified for clothing items if a trained engine is to be generated for any new dress code or any new context.

[0053] It would be advantageous if a clothing selection system could be generated that did not require such manual input, such as by not requiring explicit input regarding types of clothing that may be combinable or manually-encoded examples of combinable clothing. Where limited data can be input, a user would be able to create different clothing selection algorithms for different contexts in an easier manner, permitting usage of the training and selection system across more contexts, such as for more events, occasions, styles, etc. To enable such a system that accepts limited input, it would be advantageous if a designer could input a design theme in a more straightforward manner than in the prior art, through input of only a limited number of images or other data. In such a system, design features may be extracted from an automated analysis of those images, such as through image processing techniques. Such extracted design features may then be used to automatically define rules for clothing options that may satisfy the design theme and which may then be used to select clothing items that may be combined into an outfit in accord with the automatically-defined rules.

[0054] In some embodiments described herein, a mood board enables such input and mood board data may be analysed to extract design features that may then be

subsequently used to define a collection of design rules that are to be met by an outfit.

[0055] An automated extraction of features and automatic determination of rules may risk training a system that does not accommodate all of the different constraints on clothing and fashion that are known and were discussed above, and may not account for subjective variations between people with respect to clothing and fashion. In addition, when only limited input is provided, this imposes a great burden on the user to enter a correct input and avoid mis-training a system with poor examples. As such, for clothing/fashion, a system that automatically extracts features and determines rules may be less desirable than conventional systems that requires manual specification of rules. However, such a system may have certain advantages, as discussed above.

[0056] By utilising mood board data of a mood board generated by a designer, the initial design theme can be defined by the designer, but the user can be presented with a selection of items of clothing that are selected based on an Al algorithm, created using the mood board data. In this way, the user therefore can receive the help they would receive from a personal shopper or a friend but without the need for the designer to manually select items of clothing from the sometimes vast stock of clothing available. Consequently, the user experiences a virtual personal shopper experience.

[0057] In another example, a designer may input mood board data into a device. This device can be the user terminal 204. Alternatively, the device is a device other than the user terminal 204. For example, the device to receive input from the designer may be a different type of device to the user terminal 204. For example, the device to receive input from the designer may be a device with a dashboard, such as a computer, and the user terminal may be a mobile communication device capable of providing the user with a mobile application. In an example, the designer may input images, for example, taken on a camera of the device (not shown). The input data may include the design theme. The server 202 or user terminal 204 may receive the design theme and create the mood board data indicative of the design theme. The mood board data may be created based on the input data received at the device. It will be understood that it is not essential that the mood board data includes the design theme, as the mood board data will itself be representative of the design theme by virtue of the included images chosen by the designer.

[0058] The mood board may include one or more images that can represent a design theme. For example, images of the mood board may have been taken by a designer, for example a photographer, and the detail within the images may have been created by a designer. In an example, the designer may be a merchandiser. Features in the images can be extracted that allow for the training of the Al engine to generate a clothing item selection machine. For example, the mood board may include a significant proportion of image data of a particular colour. As illustrated in figures 3A and 3B, the mood boards 302, 352 may contain a plurality of images and at least one of these images may indicate the design theme. Typically, each of the images will be representative of the design theme. In an example, at least one image of the mood board may comprise an image of clothing. In another example, at least one image may not comprise an image of clothing. In another example, the mood board may be a set of clothing items. In another example, the mood board may contain words such as keywords related to the design theme. The words may assist the Al engine in selecting clothing items based on the design theme. For example, the mood board may contain words such as“Boho”“Texan” and“Chic”. Each mood board may be unique. Some of the features of a mood board may appear in another mood board.

[0059] The design theme is indicative of a common theme to be considered when selecting clothing items using the clothing item selection machine. For example, a user may utilise a design theme to reduce the amount of clothing items they are shown and to direct their search to appropriate clothing items which can be combined together to form an outfit corresponding to the design theme. The method of figure 1 and apparatus of figure 2A enable such a theme to be selected and implemented through the mood board. The design theme may be a type of event, type of holiday, a time of year, a particular season of clothing or a particular style of clothing. It will be understood that other design themes are possible.

[0060] In figure 3A, the mood board 302 comprises six images. In figure 3B, the mood board 352 comprises four images. In these figures, each image is of at least one model with clothing and each image has a different background. The clothing has different patterns and colours. The clothing is of a similar type. The backgrounds are of different colours and include different floorings, different walls and different objects. Moreover, the models are also different.

[0061] The clothing items 408 to 426 and 458 to 476 of figures 3A and 3B are based on the respective mood boards 302, 352, which are indicative of a design theme. Thus, each set of clothing items is based on a design theme. At least some of the clothing items selected from the database based on the mood board may be displayed to a designer on a separate device, as illustrated in figures 3A and 3B. The separate device to display to and receive input from the designer has been previously described. The designer may be the same designer as the designer of the mood board or may be a different designer. The designer may be a merchandiser. The designer may then select clothing items to be displayed to the user on the display 206 of the user terminal 204 from the clothing items displayed on the separate device. For example, the designer may be presented with a “Products” tab in the user interface. In this tab, the designer may be presented with the list of clothing items selected by the Al engine 262 and/or the clothing item selection machine 264 based on the mood board. In this example, the designer can select the clothing items to be displayed to the user by selecting the box in the corner of each clothing item as illustrated in figures 3A and 3B.

[0062] Additionally, or alternatively, the designer may be presented with an“Outfits” tab and the designer may navigate to this tab. In this tab, a designer may be presented with a list of outfits selected by the Al engine 262 and/or the clothing item selection machine 264 based on the mood board. The designer may additionally select clothing items presented in the“Products” tab to appear in the“Outfits” tab. In an example, the designer may amend an outfit to include a clothing item not selected by the Al engine. For example, the designer may include this clothing item because it may match with the other clothing items of the outfit that have been selected by the Al engine.

[0063] The designer may amend the presented outfits manually, for example, by moving clothing items between the outfits, adding clothing items into or deleting clothing items from the outfits. The outfits that appear in the“Outfits” tab may be the outfits to be displayed to a user. In an example, each of the outfits from the“Outfits” tab to be displayed to the user require approval and/or validation by the designer before they are displayed to the user. The designer’s ability to manipulate outfits enables the refinement of the outfits that will be displayed to the user. Thus, the selection of an outfit for display to a user may combine the capabilities of the Al engine 262 and/or clothing item selection machine with the capabilities of the designer. The clothing items selected in the“Products” tab and the amended“Outfits” may be used to further train the Al engine to improve its clothing item selection and combination ability.

[0064] In the“Outfits” tab of the user interface, the designer may also be able to designate alternative clothing items to be displayed to the user as an alternative to the clothing items making up the outfit. In this way, the designer can choose not just the outfits to be presented to the user, but also alternative clothing items should the user dislike one or more clothing items in the outfit.

[0065] The clothing items 408 to 426 and 458 to 476 are examples of the clothing items selected 108 from the database in method 100 of figure 1 and may be selected based on the mood boards 302, 352.As illustrated in figure 3A, the design theme for mood board 302 is“Check”. The mood board may be indicative of a design theme. In the illustration of figure 3A, the mood board contains checks to indicate that the design theme is“Check”. The clothing items 408, 410, 412, 414, 416, 418, 420, 422, 424 and 426 of figure 3A may relate to the mood board 302 and the design theme of the mood board 302. For example, clothing item 414 contains a checked dress which relates to the mood board 302 and design theme.

[0066] As illustrated in figure 3B, the design theme for mood board 302 is“Wintry Boho”. The mood board may be indicative of the design theme. In the illustration of figure 3B, the mood board contains warm clothes, for example coats, in the boho style to indicate that the design theme is“Wintry Boho”. The clothing items 458, 460, 462, 464, 466, 468, 470, 472, 474 and 476 of figure 3B may relate to the mood board 352 and the design theme of the mood board 352. For example, clothing items 460, 464, 470, 472 and 474 are coats and therefore relate to the mood board 352 and design theme. Because the mood boards 302 and 352 differ, the clothing items based on the mood boards also differ. The clothing items 408, 410, 412, 414, 416, 418, 420, 422, 424 and 426 based on the mood board 302 differ from the clothing items 458, 460, 462, 464, 466, 468, 470, 472, 474 and 476 based on the mood board 352. The description corresponding to the method 100 of figure 1 and apparatus 200 of figure 2A provides details on how the clothing items are selected given the mood board data.

[0067] Referring back to figures 1 2A, and 2B, mood board data, as described in relation to figures 3A and 3B, indicative of the design theme of the mood boards 302, 352 of figures 3A and 3B may be received 102 by the server 202. As mentioned previously, the received mood board may have been created based on user input from a device. An Al engine 262 may be included in the server 202 or may be separate from the server 202 and connected through a wired/wireless connection to the user terminal 204. The Al engine 262 is trained 104 by the server 202 based on the mood board data. The Al engine 262 receives data indicative of the design theme of the mood board and trains using this data to generate a clothing item selection machine 264 characteristic of the design theme for selecting clothing items deemed to correspond to the design theme. In an example, the Al engine 262 uses a training data set in addition to the mood board data to train. The training data set may include a small set of clothing items that represent a larger set of clothing items. For example, the training data set may include one of each type of clothing item. In an example, the Al engine 262 may be trained based on other data such as statistical data on a user’s previous purchases or user preferences such as types of clothing item or price of clothing item. In an example, a user selection input is used to further train the Al engine 262. [0068] In an example, the Al engine 262 includes a neural network, and the input of the Al engine 262 is connected to the output of the Al engine 262 through a series of connected nodes of the neural network. Thus, when the Al engine 262 receives a training set, such as clothing items, or mood board data, the connections between the nodes and the weights of the nodes themselves are modified and this is repeated until the Al engine 262 produces the desired output. The Al engine 262 may continually train based on new data received to ensure it produces the desired output. For example, the Al engine 262 may train based on user inputs such as user selection of a clothing item. It will be understood that the user input may come from one or both of the user of the user terminal on which the selected outfit is displayed rendered on the avatar, and the device used by the designer for inputting the images of the mood board making up the mood board data. For each mood board indicative of a design theme, the neural network will be different. In an example, the designer may view a plurality of clothing items produced by the Al engine 262 in order to ensure the Al engine 262 is producing the desired output. Once the Al engine 262 is trained, when receiving an input, it will be able to produce the desired output with minimal delay. For example, the delay may be so minimal that it would not be noticeable to a user.

[0069] In another example, the Al engine 262 comprises an algorithm with a set of rules that can be applied to tags of a database to find the desired clothing items. The Al engine 262 may analyse the features of the received mood board data in order to determine rules of the mood board data and consequently generate the clothing item selection machine 264. The combination of the designed mood board and the computer-implemented Al engine 262 enables the combination of the artistic capabilities of a human with the fast and efficient unsupervised learning process of the Al machine 262 to efficiently and effectively reduce the number of clothing items provided to a user.

[0070] The Al engine 262 may analyse the colours used, the types of clothing items, the style of the clothing items, the shapes used in any images of the mood board and any words present in the mood board. The Al engine 262 may use standard image processing techniques to find these features of the received mood board data. For example, an edge detection technique may be utilised on the mood board data to find any shapes or outlines of clothes in order to identify the type of clothing. Moreover, the Al engine may store or have access to databases of, for example, known shapes of types of clothing items. In this way the Al engine 262 can easily identify from edge detection and comparison with known shapes in the database the presence of a clothing item in the mood board. Where a clothing item is identified to be present, the Al engine can also categorise the type of the clothing item. It will be understood that the skilled person will be aware of a number of different image processing techniques that can be used to extract features from the images in the received mood board data.

[0071] The Al engine 262 may form a set of rules based on the mood board. For example, if the mood board data contains only red and green, the Al engine 262 may create a rule that clothing items must contain at least some red or green. Alternatively, if the mood board data contains a red skirt, the Al engine 262 may create a rule that any red skirts in the clothing items must be selected. It will be understood that the rules may be far more complicated than the simplistic examples given above. In some examples, the rules used by the Al engine may not even be readily expressible to a human observer.

[0072] The Al engine 262 is trained to generate a clothing item selection machine 264. The clothing item selection machine 264 may be for activation on the server 202. In another example, the clothing item selection machine 264 may be for execution elsewhere on the apparatus 200, separately from the server 202. In another example, the clothing item selection machine 264 may be located externally to the apparatus 200. The clothing item selection machine 264 is characteristic of the design theme because the Al engine 262 is trained based on the mood board data indicative of the design theme. In an example, the clothing item selection machine 264 may include the trained neural network of the Al engine 262. In another example, the Al engine 262 may generate an algorithm that is utilised by a clothing selection machine 264. The algorithm may include rules that the Al engine has identified. The clothing item selection machine 264 may use this algorithm to select clothing items deemed to correspond to the design theme. The clothing item selection machine 264 may identify the clothing items further based on a user selection input in relation to one or more of the clothing items in the stock of clothing. In an example, as previously mentioned, the designer may indicate which clothing items of the clothing items based on the mood board to display to the user. For example, the algorithm of the Al engine 262 may be utilised by the clothing item selection machine to produce a database of clothing items and, subsequently, the designer may then select a plurality of clothing items from the database of clothing items to display to the user. In another example, the clothing item selection machine may additionally utilise other means in order to select the clothing items, such as user input, user purchase history or user

characteristics such as age. The clothing item selection machine 264 is for selecting clothing items deemed to correspond to the design theme. By using the algorithm generated by the Al engine 262 based on the mood board data indicative of the design theme, the clothing item selection machine 264 can select items that are also indicative and correspond to the design theme.

[0073] The server 202 may access 106 a database of a stock of clothing items available through a managed supply chain. The database may be stored in memory 268 that is located in the same apparatus as the server 202. In another example, the database may be located externally to the apparatus and accessed via a wired and/or wireless connection. In an example, the database may be located with the clothing item selection machine. The database may comprise at least 1000 items of clothing. In an example, the database may receive clothing items from a plurality of merchandisers or vendors. In another example, the stock of clothing items may be sourced from one supplier such as a particular brand. The stock of clothing items may be clothing items that are available, for example in stock in a warehouse, at the time that the database is accessed. The managed supply chain may be the means in which the clothing is transferred from its original location when it is manufactured, e.g. a warehouse, to the customer. In another example, the managed supply chain may include the sourcing and supply of materials to

manufacture the clothing. In another example, the database may include clothing items previously associated with the user, for example, clothing items that a user has previously purchased. In another example, clothing items previously associated with the user may be clothing items that a user has previously saved. In another example, clothing items previously associated with the user may be clothing items that a user has previously indicated that they like, for example, by adding the clothing items to a“wish list”. In an example, the database may include clothing items previously associated with the user and may receive these clothing items from the server, which, for example, may have previously received information on the user’s previous purchases. In another example, the server may have previously received information on the user’s saved items from the user via the user terminal. In another example, the server may have previously received information on the items the user liked from the user via the user terminal. In another example, the database may receive these clothing items from a supplier that supplied the user with the previously purchased clothing items. The database of the stock of clothing items may comprise clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing. For example, the clothing characteristic data may be included in a virtual tag associated with each clothing item. Therefore, the rules of the Al engine can be easily implemented by viewing the clothing characteristic data, for example the tags, of each clothing item to determine whether each clothing item satisfies the rules and, if it does, to select that clothing item. The characteristics of each of the clothing items in the stock of clothing may include the type of clothing item, colour of clothing item, style of clothing item, season in which clothing item is available and/or cost of the clothing item.

[0074] The server 202 may then select 108 from the database a plurality of clothing items of different types. In an example, the server 202 may send a specific request to the database to send back certain clothing items. In another example, the server 202 may access the database to select the plurality of clothing items. In another example, the server 202 may receive each clothing item and only store details of certain clothing items.

[0075] The clothing items are selected by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme. For example, the clothing item selection machine 264 may determine the clothing items to select based on the tags associated with each clothing item. In an example, the server 202 may instruct the clothing item selection machine to identify the clothing items. For example, the clothing item selection machine 264 may be instructed to identify clothing items and send the identified clothing items to the server 202 which may then select the identified clothing items from the database. In another example, the clothing item selection machine may, as part of the server, access the database of the stock of clothing items and identify the clothing items for combination together as one or more outfits that correspond to the design theme. The clothing item selection machine may then inform the server 202 or the interface between the server 202 and the database which then selects the plurality of clothing items of different types. In another example, the clothing item selection machine may identify and select from the database, based on the clothing characteristic data of the database, the clothing items for combination together as one or more outfits that correspond to the design theme. In an example, the server 202 may generate, using the clothing item selection machine 264, one or more outfits that correspond to the design theme as a combination of a plurality of selected clothing items of different types, the outfits being for displaying on the display of a user terminal 204.

[0076] The server 202 may receive user characteristic data indicative of one or more user characteristics including user size characteristics. The server may receive the user characteristic data from the user terminal 204. For example, the user characteristic data may be input by the user. In another example, the server may receive the user

characteristic data from another server or terminal that stores data about multiple users. In another example, the server may receive the user characteristic data from a database within the server that stores data about multiple users. For example, the server may receive user size characteristics including at least one of waist, chest or height. In another example, the server may receive the clothing size of a user from a user input on the user terminal 204 indicating their clothing size or from a purchase history of the user. For example, the user may purchase clothing items of a selected size and the size

characteristics may be based on the type of clothing item and the selected size that the user has purchased. When the server 202 selects 108 from the database a plurality of clothing items, this selection may also be based on the user characteristic data. Moreover, selecting 108 the plurality of clothing items from the database may also be based on inventory of deliverable stock of the clothing items matching the user characteristic data.

[0077] As previously mentioned, the clothing item selection machine 264 may receive an algorithm from the Al engine 262 to select clothing items deemed to correspond to the design theme and the algorithm may contain rules that the Al engine has identified. The clothing item selection machine may therefore compare these rules to the clothing characteristic data of the database. For example, if a rule of the algorithm is a red skirt, the clothing item selection machine may search the clothing characteristic data of the database to find the words“red” or“skirt”, for similar characteristics such as“maroon” or for other characteristic data that corresponds to such words. In an example, the clothing item selection machine identifies a plurality of clothing items and the server 202 then selects one or more clothing items from the identified plurality of clothing items in order to produce an outfit. In another example, the server 202 selects a plurality of clothing items from the database and, for each clothing item, creates an outfit using other selected clothing items from the database or clothing items from the database that have not been selected.

[0078] The clothing item selection machine 264 may also be provided with an algorithm from the Al engine or information from an alternative source that enables the clothing item selection machine to identify other clothing items to combine with the clothing item to produce one or more outfits that correspond to the design theme. In an example, the clothing item selection machine may be provided with rules on which clothing items to combine, including which types of clothing item to combine, in order to produce an outfit. For example, the machine could be provided with information on the types of clothing items that can be combined for each outfit. In another example, the clothing item selection machine may receive information on which types of clothing items to select based on other types of clothing items selected. For example, the clothing item selection machine may select a pair of trousers and receives information that it must also select a top. In an example, the clothing item selection machine may also receive information on clothes matching to assist with combining clothing items to create an outfit. In another example, the clothing item selection machine may be provided with a number that specifies the amount of clothing items the clothing item selection machine can select.

[0079] The server 202 then locates 1 10 clothing image data representative of the selected clothing items. For at least some of the types of clothing, this clothing image data is representative of the clothing items as they would appear when worn by a wearer in a predetermined pose. In an example, the image data is located within and received from the database. In another example, the image data is received from a second database different from the database. The server 202 may send information to the database or the second database to ensure the correct image data is located and received. In another example, the server may not receive the clothing image data but may locate the clothing image data which may then be transmitted to and received by the user terminal 204. For each clothing item selected from the database, the server 202 may locate the

corresponding clothing image data representative of the clothing item. In an example, the server 202 compares the clothing item to the clothing image data to locate the correct clothing image data. In another example, the clothing image data may be stored with or linked to the clothing item and the server 202 may therefore not have to perform a comparison step.

[0080] The clothing image data is used to form an image of the clothing item that it represents. In an example, the clothing image data, when displayed as an image, comprises the clothing item as if it was being worn by a person. For example, the clothing image data may comprise an image of a clothing item on a person in a predetermined pose without comprising the person. In another example, each clothing item may be represented by multiple images of the clothing item as if it was on a person, with each image showing the clothing for a person in a different pose. Thus, the clothing image data may be used to produce multiple images. In another example, the image created from the clothing image data is a 3-dimensional image that can be rotated and/or manipulated to illustrate the clothing in different positions. In this example, the clothing image data is representative of the clothing items as they would appear when worn by a wearer in a plurality of poses due to the positions available. This enables flexibility in the poses of models and in the views of the clothing available to a user. In a shop, a user may pick clothing up and view the clothing from a plurality of positions and may try the clothing on and view the clothing from a plurality of positions. Thus, by providing clothing image data that represents a plurality of poses of a person, the server 202 enables the user to receive the same experience online as in a shop. By enabling views of the clothing from a plurality of positions, the user can view the clothing in a way that more closely matches their experience in a shop when they’re able to view the physical clothing rather than an image, and the user is therefore provided with a more realistic experience. This allows the user to obtain more details on clothing items to assist with their selection.

[0081] The server 202 then causes 1 12 display data to be transmitted to a user terminal 204. The display data may include the located clothing image data. The display data may include the one or more clothing items selected by the server 202 to be displayed. The server 202 may transmit the display data to the user terminal 204. In another example, the server 202 may instruct an external database storing the clothing image data to transmit the data to the user terminal 204. For example, the server 202 may instruct the database or the second database to transmit the clothing image data to the user terminal 204. [0082] The user terminal 204 may receive the display data. In response to receiving the display data from the server 202, the display 206 of the user terminal 204 then displays 112, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar. In an example, the avatar is representative of a real person. In an example, the real person is a model. In another example, the real person is the user of the user terminal 204. The avatar may be a picture of a person in a predetermined pose. In another example, the avatar is substantially photo-realistic, such that the representation of the at least one item of clothing on the avatar is suitable to illustrate how the item of clothing would appear on a person. In another example, the avatar is a 3-dimensional image.

[0083] In an example, the image data representative of an avatar of a person is received from the server 202. In another example, the image data representative of an avatar of a person is stored on the user terminal 204. In another example, the image data

representative of an avatar of a person is created by the user of the user terminal 204. Thus, the image data representative of an avatar of a person may be unique to the user. In another example, the image data representative of an avatar of a person is image data of the avatar in a predetermined pose. In another example, the image data representative of an avatar of a person includes image data of the avatar in a plurality of poses.

[0084] The outfit may consist of a combination of a plurality of selected clothing items.

For example, the outfit may comprise five clothing items. The outfit may consist of at least two different types of clothing items. The types of clothing items that form an outfit may be predetermined. For example, the outfit may comprise a dress, shoes and a bag. In another example, the outfit may comprise a skirt, a top, earrings and shoes. In an example, the amount of clothing items is not constant for all outfits but differs depending on the outfit type. In an example, the plurality of clothing items are combined to form an outfit when selected from the database. In another example, the user terminal 204 combines the clothing items for display as an outfit.

[0085] At least one clothing item of the outfit is rendered as being worn by the avatar.

The rendering the at least one of the plurality of clothing items on the avatar may comprise combining the representations of the plurality of clothing items with a representation of the avatar, such that the avatar appears to be wearing the at least one clothing item. In an example, at least one clothing item of the outfit is not worn by the avatar. In an example, the image of the clothing item may be selected based on the pose of the avatar. In an example, if the image is 3-dimensional, the image may be manipulated to ensure the positioning of the clothing item is at least similar to the pose of the avatar. In another example, the avatar and the image may both be 3-dimensional. If both the avatar and the image are 3-dimensional, the one or more clothing items may be rendered as being worn by the avatar such that a user can manipulate the avatar to view the clothing items on the avatar in different poses. In this example, a user may be able to fully rotate the avatar whilst still viewing the clothing items as if they are being worn by the avatar.

[0086] The image of the clothing item should be in substantially the same positioning as the pose of the avatar such that when the user views the display, the user believes that the avatar is wearing the clothing items. This enables a user to view clothes online as if they or a friend were to wear the clothes in the shop. The one or more clothing items could be rendered as being worn by the avatar by combining the images of the clothing items and avatar before displaying 112 on the display 206 of the user terminal 204. In another example, the one or more clothing items could be rendered on the avatar by producing the avatar in a layer and the clothing in another layer in front of the avatar layer on the screen. In another example, displaying 112 the image data comprises combining the image data representative of the avatar with the image data representative of the at least one clothing item and geometrically transforming the image of the or each at least one clothing item of the plurality of clothing items to appear fitted to the representation of the person, and rendering the combined image data. For example, the image of the clothing item may be stretched to appropriate proportions or rotated in order to be effectively rendered as being worn by the avatar. The display may also display 1 12 an indication of the design theme with which the outfit is associated in juxtaposition to the avatar and the outfit, as will be described later with reference to figure 4A.

[0087] In an example, the user terminal 204 may render the one or more clothing items on the correct part of the avatar by using image processing means such as edge detection. In an example, the user terminal 204 may select the image data of the clothing item to use based on the image processing of the avatar. In another example, the pose of the avatar may be constant such that only one image of the clothing image data is necessary. In another example, the pose of the avatar may be included in the characteristics of the avatar and the clothing characteristic data of the clothing items may contain the pose of the clothing item such that the two poses can be compared to ensure the correct pose of clothing is matched with the correct pose of the avatar.

[0088] The display may also be caused to display image data representative of at least one further outfit combining a plurality of the selected clothing items, wherein at least one clothing item associated with each further outfit is rendered as being worn by the avatar. In an example, the further outfit may be displayed on the display but may not be displayed on an avatar. In another example, at least one clothing item of the further outfit may be displayed on the display but not on the avatar. In an example, the further outfit comprises at least one retained clothing item also present in the first outfit, and a further clothing item different to any of the clothing items associated with the first outfit. The representation of each of the plurality of clothing items associated with the second outfit may be caused to be displayed on the electronic display in response to a change request received from the user and the change request may be indicative of a request from the user to change a clothing item in the first outfit.

[0089] In an example, the display may display more than one avatar, wherein a different outfit is rendered as being worn on each avatar. Figures 4A, 4B, 5A, 5B, 6A, 6B, 7A and 7B further detail examples of how the clothing items and avatar are displayed on the display 206 of the user terminal 204.

[0090] Figures 4A and 4B provide schematic examples 400, 450 of the display 206 of the user terminal 204. These figures provide examples of the method step 112 of figure 1. On a user terminal 204, a user may navigate from the home screen of the user terminal to a user application by clicking on an icon. In response to the user input, the user terminal 204 may display the home screen of the application. The user terminal 204 may receive a user input indicating their desire to view a particular design theme 402.

[0091] In an example, the server 202 has already trained the Al engine for the particular design theme and transmitted the relevant display data to the user terminal 204 for the user terminal 204 to display. For example, one or more design themes may automatically be utilised to perform the method 100 of figure 1. For example, each day, certain design themes may be predetermined and the method of figure 1 performed. This enables a user to discover a new set of curated outfits grouped by design themes each day. These design themes may be indicated to the user on the home page of the application. Thus, the user terminal 204 may already contain the clothing image data for the clothing items and may display these to the user in response to the user input indicating their desire to view the particular design theme 402, for example, as illustrated in figure 4A.

[0092] In an example, design themes may be decided for the next few months or year. The designers may then design the mood boards based on the design themes in advance of the mood boards being utilised by the Al engine.

[0093] Alternatively, the Al engine may not have been trained for the selected design theme. For example, the user may select a design theme that is not in the predetermined list for that day. Thus, the design theme may be transmitted to the server 202 which performs the method steps as explained with reference to figures 1 , 2A and 2B. After the clothing image data for the clothing items has been received, the user terminal 204 displays these to the user in response to the user input indicating their desire to view the particular design theme 402, for example, as illustrated in figure 4A.

[0094] Image data representative of a series of outfits of selected clothing items may be displayed sequentially on the display of the user terminal 204. Figure 4A provides an example of a first display 400 displaying a first outfit 406. When a predetermined time has passed, as selected by the user terminal 204 or the user, or a user has interacted with the user terminal 204 such that a user input has been recorded, the first display 400 of figure 4A transitions to the second display 450 of figure 4B which displays a second outfit 456.

[0095] Referring to figure 4A, the display 206 is configured to display, in response receipt from the server 202 of display data, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar. In an example, in figure 4A the display 206 displays a plurality of clothing items 408, 410 and 412. The display 206 may display an outfit 406, that comprises clothing items 408 and 410. Display 206 may also display an avatar 404. Clothing item 408 is rendered as being worn by the avatar. In an example, other clothing items and/or an outfit may be rendered as being worn by the avatar. In an example, as illustrated in figure 4A, clothing item 408 and clothing item 410 of outfit 406 may be rendered as being worn by the avatar. In another example, only clothing item 408 may be rendered as being worn by the avatar, for example, clothing item 408 may be a dress. In an example, outfit 406 may comprise clothing item 412 but clothing item 412 may not be rendered on avatar 404 and may be located separately to the rest of the outfit 406. For example, clothing item 412 may be located below the avatar 404 (as illustrated in figure 4A) or adjacent to the avatar 404. For example, at least one clothing item of the outfit may not be rendered on the avatar. In an example, at least one of the plurality of clothing items is displayed on the electronic display of the user terminal 204 off the avatar and adjacent to the avatar. For example, the clothing items, for example clothing item 412, displayed adjacent to the avatar comprises at least one of: footwear, a bag, jewellery and a wallet and a purse. In another example, outfit 406 may not comprise clothing item 412. The dashed line in this figure demonstrates that outfit 406 may include clothing item 412 and clothing item 412 may or may not be rendered as being worn by the avatar. Display 206 may also display design theme 402.

[0096] In an example, the illustrations of figure 4A and 4B are frames of a video, and the video displays the avatar and outfit with a moving background that relates to the design theme of the outfit. This video may be of a predetermined length, for example 5 seconds. Once the video of figure 4A has finished, the display may then transition to the video of figure 4B.

[0097] Referring now to Figure 4B, in this example, display 206 of user terminal 204 displays design theme 452, avatar 454, outfit 456 and clothing items 458, 460 and 462. These modules displayed on display 206 may be arranged as disclosed in relation to figure 4A. In an example, outfit 456 is a different outfit to outfit 406 of figure 4A. In an example outfit 456 comprises clothing items 458, 460, 462 that are different to clothing items 408, 410, 412. Clothing items 458, 460, 462 may be the same type as clothing items 408, 410, 412 but may differ in other ways, for example, brand, colour, shape and material. Avatar 454 may be the same as avatar 404. In an example, avatar 454 stays the same during the transition from figure 4A to figure 4B but the outfit rendered as being worn by the avatar is changed from outfit 406 to outfit 456. In an example, design theme 452 is different to design theme 402. Thus outfit 456 is based on a different design theme, and thus different mood board, to outfit 406. In another example, design theme 452 is the same as design theme 402 and outfit 406 and outfit 456 differ but are both included in the plurality of selected clothing items selected from the database of the stock of clothing items. Oufits 406 and 456 may comprise a different amount of clothing items. For example, in figure 4A, outfit 406 may comprise clothing items 408, 410 and 412 but, in figure 4B, outfit 456 may comprise clothing items 458 and 460 but may not comprise clothing item 462. In another example, outfit 456 comprises clothing items 458, 460 and 462 and clothing items 458 and 460 are rendered as being worn by the avatar 454. Clothing item 462 may be located below the avatar 454 (as illustrated in figure 4B) or adjacent to the avatar 454.

[0098] Figures 5A and 5B provide illustrative examples of a display of a user terminal 204. The illustration 500 of display 206 at figure 5A is an example of the schematic example 400 of display 206 of figure 4A and the illustration 550 of display 206 at figure 5B is an example of the schematic example 450 of the display 206 of figure 4B. As can be seen, the illustration 500 of figure 5A shows an outfit comprising a dress, a bag and shoes. In this example, the outfit comprises a dress as clothing item 408, shoes as clothing item 412 and earrings as clothing item 416. In this example, only the dress, clothing item 408, is rendered on the avatar 404. The illustration 550 of figure 5B shows an outfit comprising a dress, earrings and shoes. In this example, the outfit comprises a dress as clothing item 458, shoes as clothing item 462 and a bag as clothing item 466. In this example, only the dress, clothing item 458, is rendered on the avatar 454. In this example, each of the items of clothing in the outfits in illustrations 500, 550 are different but the design themes 402, 452 are the same.

[0099] Figures 6A and 6B provide further schematic examples of a display 206 of a user terminal 204. In an example, figures 6A and 6B illustrate the method step 112 of figure 1. Alternatively, during the display of the image data, for example, as illustrated by figures 4A, 4B, 5A and 5B, the user may interact with the user terminal. For example, the user may indicate that they are interested in an outfit displayed on the display, for example, outfit 406. When the display is showing outfit 406 as illustrated in figure 4A, the user may provide a user input, for example the user may tap the display or gesture on the display, for example swipe up. This may indicate that the user is interested in the outfit. The display may then transition to show a different illustration of the outfit. For example, if the display is displaying the illustration of figure 4A and the user indicates that they are interested in the outfit, the display may then display the illustration of outfit 406 as shown in figure 6A.

[00100] Figure 6A provides an example of a first display 600. When a predetermined time has passed, as selected by the user terminal 204 or the user, or a user has interacted with the user terminal 204 such that a user input has been recorded, the first display 600 of figure 6A transitions to the second display 650 of figure 6B. In an example, a user may desire to view clothing item 458 on the avatar 404. Therefore, the user may interact with the terminal to indicate that the user has chosen the selected clothing item 458. Thus, at least one of the clothing items of the outfit 406 may be changed based on the user input. In an example, when the user selects clothing item 458 in figure 6A, the display 206 of the user terminal 204 then transitions to include the selected clothing item 458 in the outfit and renders the clothing item on the avatar or adjacent to or below the avatar. The user terminal 204 itself may combine the selected clothing item 458 with the rest of the outfit. The user terminal 204 itself may render the clothing item and/or outfit as being worn by the avatar. The user terminal 204 may then display the selected clothing item 458 in the outfit and, in some examples, rendered as being worn by the avatar. In another example, the user terminal 204 may transmit the clothing item to a server 202 and the server 202 may combine the selected clothing item 458 with the outfit and/or render the clothing item 458 as being worn by the avatar. The server 202 may subsequently transmit the display data including the outfit and/or avatar to the user terminal 204 for the user terminal 204 to display. In an example, the user may touch the touch screen or click a cursor located near clothing item 458. In response to the user input on the user terminal 204, image data representative of the avatar and of an adapted outfit including the user-chosen clothing item may be displayed on the display 206. As illustrated in figure 6B, the user selected clothing item may then appear as part of outfit 456 in figure 6B and may be rendered on avatar 454 of figure 6B. This user input may initiate the transition between the displays shown in figure 6A and figure 6B. For example, when the display 600 of figure 6A is displayed and the user indicates that they are interested in clothing item 458, which does not appear on the avatar, the display transitions to the display 650 of figure 6B, wherein the avatar is now displayed with the outfit 456 containing the clothing item 458 in which the user is interested. The user inputs, including the clothing items the user selects, may be used to further train the Al engine such that the Al engine is unique to the user of the user terminal 204. It will be understood that in examples only one of the items of clothing shown in the avatar 404 may be changed at a time.

[00101] Referring to figure 6A, the display 206 is configured to display, in response receipt from the server 202 of display data, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar. In an example, in figure 6A the display 206 displays a plurality of selected clothing items 408, 410, 412, 414 and 458. The display 206 may display an outfit 406, that comprises clothing items 408 and 412. In an example, outfit 406 does not comprise clothing items 414 and 458 and these clothing items are displayed in a separate location to the outfit 406. Display 206 may also display an avatar 404. Clothing item 408 may be rendered as being worn by the avatar. In an example, other clothing items and/or an outfit may be rendered as being worn by the avatar. In an example, as illustrated in figure 4A, clothing item 408 and clothing item 410 of outfit 406 may be rendered as being worn by the avatar. In an example outfit 406 may comprise clothing item 412 but clothing item 412 may not be rendered on avatar 404. For example, at least one clothing item of the outfit 406 may not be rendered on the avatar. The dashed line in this figure demonstrates that outfit 406 may include clothing item 412 but clothing item 412 may or may not be rendered as being worn by the avatar. For example, clothing item 412 may be located below the avatar 404 or adjacent to the avatar 404 (as illustrated in figure 6A).

[00102] In an example, clothing items 414, 458 may not be part of outfit 406. Therefore, selected clothing items 414, 458 may be displayed in a separate location to outfit 406. For example, selected clothing items 414, 458 may be displayed in juxtaposition to the avatar and the outfit. The selected clothing items 414 and 458 may be suggested clothing items that have been provided based on the clothing items 408, 410 in outfit 406. For example, the clothing items 414, 458 may be suggested clothing items to replace clothing items 408, 410 in outfit 406. In an example, clothing items 414, 458 may be clothing items selected by the clothing item selection machine.

[00103] Referring to figure 6B, in this example, the display 206 displays avatar 454, outfit 456 and clothing items 458, 460, 462, 408 and 412, which may be arranged in the same way as the display of figure 6A, as described previously. In an example, avatar 454 of figure 6B may be the same avatar as avatar 404 of figure 6A. Outfit 456 may be a different outfit to outfit 406 and may comprise different clothing items. In another example, clothing items 458 and 462 of outfit 456 may be different to clothing items 408 and 412 of outfit 406 but clothing item 460 of outfit 456 may be the same as clothing item 410 of outfit 406. For example, a further outfit 456 may comprise at least one retained clothing item 410 also present in the first outfit 406, and a further clothing item 458 different to any of the clothing items associated with the first outfit 406, and the representation of each of the plurality of clothing items associated with the second outfit 456 is caused to be displayed on the electronic display 206 in response to a user input indicating a change request, wherein the change request is indicative of a request from the user to change a clothing item in the first outfit. In another example, outfit 456 comprises clothing items 458, 460 and 462 and clothing items 458 and 460 are rendered as being worn by the avatar 454. Clothing item 462 may be located below the avatar 454 or adjacent to the avatar 454 (as illustrated in figure 6B).

[00104] By enabling the user to select the clothing items to view on the avatar, the user is able to view their desired combination on the avatar and it is therefore easier for the user to select the clothing items and/or outfit that they wish to buy. The user is also more comfortable with their decision. Once a user has decided on their desired outfit

combination, they can share the image with other people. For example, the user can share by message, Whatsapp (RTM), Instagram (RTM), Twitter (RTM), Pinterest (RTM) or Facebook (RTM).

[00105] Figures 7 A and 7B provide illustrative examples of a display of a user terminal 204. The illustration 700 of display 206 at figure 7A is an example of the schematic example 600 of the display 206 of figure 6A and the illustration 750 of display 206 at figure 7B is an example of the schematic example 650 of the display 206 of figure 6B.

[00106] In an example, a user may be viewing the display 206 of figure 5A. The user may decide that they like the outfit and may therefore interact with the user terminal. For example, when the outfit is being displayed, the user may swipe up to view the outfit in a detailed view, such as the view illustrated in figure 7A. Clothing items 408, 412 and 416 of illustration 700 of figure 7A may be the same as clothing items 408, 412 and 416 of illustration 500 of figure 5A. Some clothing items displayed in figures 7A and 7B may not be part of the outfit displayed in the figures. For example, in figure 7A, clothing items 458 and 414 are displayed but are not part of the outfit displayed in the figure. When an outfit is first displayed, a clothing item of the outfit may be highlighted to the user. For example, in figures 7A and 7B, the dresses that are part of the outfit and displayed on the avatar are displayed in the centre of the other dresses and have a box around them.

[00107] Once the user is presented with the illustration 700 of figure 7A, the user can change the displayed outfit by replacing the clothing items of the outfit with the displayed clothing items that are not part of the outfit. The user may navigate through the different types of clothing items to find clothing items that they want to include in the outfit. The user can select clothing items to be included in the outfit and displayed on the avatar. In an example, when the display 206 is displaying the illustration 700 of figure 7A, the user may select clothing item 458. The display may then transition to illustration 750 of figure 7B, where the selected clothing item 458 is included as part of the outfit displayed in this figure and is rendered as being worn by the avatar.

[00108] For example, the user may select the tab“dresses” and then scroll through the dresses, for example, by swiping left or right. Once the user finds a dress they like, they can select the dress 458, for example, by touching the section of the display 206 displaying the dress 458. The selection of the dress 458 may cause the current dress 408 rendered on the avatar to be removed and replaced with the selected dress 458, whilst other clothing items such as the earrings 416 and shoes 412 may remain the same, as illustrated in the transition from figure 7 A to figure 7B. Alternatively, other clothing items of the outfit may change based on the selected clothing item. For example, clothing items may be chosen as part of the outfit to combine with the selected clothing item. In another example, instead of a dress being rendered on the avatar, other types and amounts of clothing items that form an outfit may be rendered on the avatar, for example, a t-shirt and trousers. The user selection of a t-shirt may then only replace the t-shirt rendered on the avatar and the trousers rendered on the avatar may remain the same. Where the avatar is shown having a dress rendered thereon, and the user selects a different type of clothing for only a portion of the avatar previously covered by the dress, for example a t-shirt, another clothing item may be automatically selected to cover the remainder of the avatar, for example a pair of trousers.

[00109] The user can also choose to remove clothing items, for example, the earrings 416, as illustrated in the transition from figure 7A to figure 7B. Figure 7B illustrates a box labelled“CHOOSE BAG” which is used as a placeholder for a clothing item 466 in figure 7B where the clothing item has been removed or the outfit does comprise many clothing items. The box may then allow the user to select another clothing item as part of the outfit, and the box may specify the type of clothing item to select. In the example of figure 7B, the box indicates that the user should select a bag.

[00110] The user can select different types of clothing items and can select clothing items to be included in the outfit but not on the avatar. For example, the user may click on the tab“Bags & Accessories” which will then cause the display to display the different clothing type of bags and accessories. The user can view the different bags and accessories by scrolling through these, for example, by swiping left or right. The user may select a clothing item from the displayed bags and accessories, for example, by touching the section of the display 206 displaying the clothing item. If the user has removed a clothing item, for example the earrings, the selected clothing item will appear on the display where the clothing item has been removed.

[00111] In another example, the user may click on the tab“Shoes” which will then cause the display to display the different clothing type of shoes. The user can view the different shoes by scrolling through these, for example, by swiping left or right. The user may select a clothing item from the displayed shoes, for example, by touching the section of the display 206 displaying the clothing item. If the user has not removed a clothing item, a clothing item may be automatically removed in order to display the selected clothing item as part of the outfit. The clothing item to be removed may be the same type of clothing item as the selected clothing item. For example, the shoes currently displayed may be replaced by the selected shoes.

[00112] Throughout the description and claims of this specification, the words“comprise” and“contain” and variations of them mean“including but not limited to”, and they are not intended to (and do not) exclude other moieties, additives, components, integers or steps. Throughout the description and claims of this specification, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating plurality as well as singularity, unless the context requires otherwise.

[00113] Features, integers, characteristics, compounds, chemical moieties or groups described in conjunction with a particular aspect, embodiment or example of the invention are to be understood to be applicable to any other aspect, embodiment or example described herein unless incompatible therewith. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and/or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and/or steps are mutually exclusive. The invention is not restricted to the details of any foregoing embodiments.

The invention extends to any novel one, or any novel combination, of the features disclosed in this specification (including any accompanying claims, abstract and drawings), or to any novel one, or any novel combination, of the steps of any method or process so disclosed.

[00114] Further examples of the present disclosure are set out in the following numbered clauses, which specify subject matter that may form the basis of claims.

[00115] 1. A computer-implemented method of providing access to a stock of clothing, the method comprising:

receiving a plurality of images indicative of a design theme;

training an artificial intelligence engine to select clothing items to be consistent with the design theme, the training comprising:

extracting, with an artificial intelligence engine, at least one design feature from the plurality of images, wherein extracting the at least one design feature comprises, for each image of the plurality of images:

identifying one or more colours and/or materials and/or patterns of an object depicted in an image as one or more design features; and

recognizing an object depicted in an image and identifying the object and/or one or more characteristics of the object as one or more design features; analysing, with the artificial intelligence engine, the at least one design feature extracted from the plurality of images indicative of the design theme; and

establishing, with the artificial intelligence engine, a collection of design rules based on the analysing of the at least one design feature;

selecting, with the artificial intelligence engine trained in the training, at least one proposed outfit to be consistent with the design theme, wherein selecting the at least one proposed outfit comprises:

evaluating, with the artificial intelligence engine and the collection of design rules, a stock of clothing items using the collection of design rules to identify a collection of clothing items consistent with the design theme;

selecting, from the collection of clothing items, at least one proposed outfit consistent with the design theme, each proposed outfit comprising a plurality of clothing items from the collection, the plurality of clothing items of each proposed outfit comprising clothing items of a plurality of clothing types;

rendering image of a person wearing a proposed outfit of the at least one proposed outfit, wherein rendering the three-dimensional image of the human comprises:

obtaining clothing image data for each clothing item of the proposed outfit being rendered, the clothing image data for at least some of the clothing items of the proposed outfit being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose;

obtaining image data for an avatar of the person; and

rendering, based on the clothing image data and the image data for the avatar, image data representative of the avatar of the person in a predetermined pose and of the proposed outfit being worn by the avatar.

[00116] 2. The computer-implemented method of clause 1 , wherein:

recognizing an object depicted in an image and identifying the object and/or one or more characteristics of the object as one or more design features comprises identifying a type of clothing item depicted in an image and identifying the type of the clothing item as a design feature;

identifying the type of clothing item comprises:

performing an edge detection technique on the image to identify at least one shape or outline in the image; and comparing the at least one shape or outline to known shapes of clothing items to identify the type of clothing item depicted in the image.

[00117] 3. The computer-implemented method of clause 2, wherein recognizing the object depicted in the image and identifying the object and/or one or more characteristics of the object as one or more design features comprises identifying a style of the clothing item depicted in the image and identifying the style of the clothing item as a design feature; and

identifying the style of clothing item comprises:

performing an edge detection technique on the image to identify at least one shape or outline in the image; and

comparing the at least one shape or outline to known shapes of styles of clothing items to identify the style of clothing item depicted in the image.

[00118] 4. The computer-implemented method of clause 3, wherein establishing the collection of design rules comprises establishing a rule for each of the at least one design feature, the rules for the at least one design feature indicating that an outfit consistent with the design theme is to meet each of the rules.

[00119] 5. The computer-implemented method of claim 1 , wherein:

receiving the plurality of images indicative of the design theme comprises receiving a plurality of images of clothing items; and

extracting the at least one design feature comprises performing at least one image processing technique on the plurality of images of clothing items to extract one or more design features.

Claims

1. A computer-implemented method of providing access to a stock of clothing, the method comprising:

receiving mood board data indicative of a design theme of a mood board;

training an Artificial Intelligence, Al, engine based on the mood board data to generate a clothing item selection machine that is characteristic of the design theme for selecting clothing items deemed to correspond to the design theme;

accessing a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing;

selecting from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme;

locating clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a predetermined pose; and causing to be displayed on a display of a user terminal image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

2. The method of claim 1 , wherein the mood board data comprises at least one image indicative of the design theme.

3. The method of claim 2, wherein the at least one image comprises an image of clothing.

4. The method of claim 2 or claim 3, wherein the at least one image comprises an image not of clothing.

5. The method of any preceding claim, further comprising generating, using the clothing item selection machine, one or more outfits that correspond to the design theme as a combination of a plurality of selected clothing items of different types, the outfits being for displaying on the display of a user terminal.

6. The method of any preceding claim, further comprising causing to be displayed on the display of the user terminal image data representative of at least one further outfit combining a plurality of the selected clothing items, wherein at least one clothing item associated with each further outfit is rendered as being worn by the avatar.

7. The method of any preceding claim, further comprising causing to be displayed sequentially on the display of the user terminal image data representative of a series of outfits of selected clothing.

8. The method of any preceding claim, further comprising:

causing to be displayed on a display of a user terminal in juxtaposition to the avatar and the outfit, other selected clothing items; and

causing to be displayed on a display of a user terminal, in response to user input on the user terminal in relation to choosing one of the displayed other selected clothing items, image data representative of the avatar and of an adapted outfit including the user-chosen clothing item.

9. The method of any preceding claim, wherein the avatar is substantially photo- realistic, such that the representation of the at least one item of clothing on the avatar is suitable to illustrate how the item of clothing would appear on a person.

10. The method of claim 9, wherein the avatar is representative of a real person.

1 1. The method of claim 10, wherein the real person is the user of the user terminal.

12. The method of any preceding claim, wherein rendering the at least one of the plurality of clothing items on the avatar comprises combining the representations of the plurality of clothing items with a representation of the avatar, such that the avatar appears to be wearing the at least one clothing item.

13. The method of claim 12, wherein causing to be displayed on a display of a user terminal image data comprises combining the image data representative of the avatar with the image data representative and geometrically transforming the image of the or each at least one clothing item of the plurality of clothing items to appear fitted to the

representation of the person, and rendering the combined image data.

14. The method of any preceding claim, further comprising causing to be displayed on a display of a user terminal in juxtaposition to the avatar and the outfit an indication of the design theme with which the outfit is associated.

15. The method of any preceding claim, wherein at least one of the plurality of clothing items is displayed on the electronic display of the user terminal off the avatar, adjacent to the avatar.

16. The method of claim 15, wherein the at least one of the plurality of clothing items displayed adjacent to the avatar comprises at least one of: footwear, a bag, jewellery and a wallet and a purse.

17. The method of any preceding claim, comprising creating mood board data indicative of the design theme.

18. The method of claim 17, wherein the mood board data is created based on one or more data inputs received at a device other than the user terminal.

19. The method of any preceding claim, wherein the clothing item selection machine identifies the clothing items further based on a user selection input in relation to one or more of the clothing items in the stock of clothing.

20. The method of claim 19, wherein the user selection input is used to further train the Al engine.

21. The method of any preceding claim, further comprising receiving user characteristic data indicative of one or more user characteristics including user size characteristics, and wherein selecting from the database a plurality of clothing items is further based on the user characteristic data.

22. The method of claim 21 , wherein selecting from the database a plurality of clothing items is further based on inventory of deliverable stock of the clothing items matching the user characteristic data.

23. The method of claim 6, wherein the further outfit comprises at least one retained clothing item also present in the first outfit, and a further clothing item different to any of the clothing items associated with the first outfit, and wherein the representation of each of the plurality of clothing items associated with the second outfit is caused to be displayed on the electronic display in response to a change request received from the user, wherein the change request is indicative of a request from the user to change a clothing item in the first outfit.

24. The method of any preceding claim, wherein the method further comprises selecting the outfit from the plurality of clothing items of different types identified by the clothing item selection machine.

25. The method of any preceding claim, wherein the plurality of clothing items of different types comprises a respective plurality of clothing items for each of the different types of clothing items.

26. A system of apparatus for providing access to a stock of clothing, the system comprising:

a server configured to:

receive mood board data indicative of a design theme of a mood board; train an Artificial Intelligence, Al, engine based on the mood board data to generate a clothing item selection machine characteristic of the design theme for selecting clothing items deemed to correspond to the design theme;

access a database of a stock of clothing items available through a managed supply chain, the database comprising clothing characteristic data indicative of one or more characteristics of each of the clothing items in the stock of clothing;

select from the database a plurality of clothing items of different types by using the clothing item selection machine to identify, based on the clothing characteristic data, clothing items for combination together as one or more outfits that correspond to the design theme;

locate clothing image data representative of the selected clothing items, the clothing image data for at least some of the types of clothing being representative of the clothing items as they would appear when worn by a wearer in a

predetermined pose; and

cause display data to be transmitted to a user terminal;

and a user terminal configured to:

display, on a display of the user terminal and in response receipt from the server of display data, image data representative of an avatar of a person in a predetermined pose and of an outfit combining a plurality of the selected clothing items, wherein at least one clothing item is rendered as being worn by the avatar.

27. A server configured as recited in the system claimed in claim 26.

28. A user terminal configured as recited in the system claimed in claim 26.