Back to results
US 11,538,083 B2Granted
Cognitive fashion product recommendation system, computer program product, and method
AssigneeInternational Business Machines Corporation
InventorsMohit Sewak, Iman Choudhury
Life of this patent
Estimating the term…
Abstract
A method, computer program product, and computing system for associating one or more fashion products on a website with a user accessing the website. One or more recommendations may be provided to the user for fashion products based upon, at least in part, one or more fashion-ability scores representative of the one or more fashion products associated with the user on the website and one or more fashion-ability scores representative of the one or more fashion products on the website.
Claims20
A computer-implemented method, executed on a computing device, comprising:
- training, according to a process, a plurality of neural networks from a plurality of images, wherein each neural network corresponds to an attribute of the plurality of images, a category of the plurality of images, or a sub-category of the plurality of images, wherein the process is selected and identified to process one or more images of a fashion category and/or a fashion sub-category to generate one or more fashion-ability scores each comprising a numerical representation of a corresponding fashion product defined in a corresponding image of the corresponding fashion product;
- for each trained neural network, scoring each of the plurality of images against each attribute that the neural network is trained for, and generating one or more scored vectors wherein each scored vector corresponds to a particular attribute used to train the neural network;
- joining each of the scored vectors for each of the plurality of images to form a multi-dimensional vector corresponding to a visual representation of each of the plurality of images;
- associating, at the computing device, one or more fashion products on a website with a user accessing the website, wherein each of the one or more fashion products associated with the user on the website is assigned a priority score based upon, at least in part a fashion browsing sequence of the user;
- generating the one or more fashion-ability scores such that the one or more fashion-ability scores are representative of the one or more fashion products on the website, by selecting an attribute and retrieving the scored vector that is trained for the selected attribute;
- in response to retrieving the scored vector trained for the selected attribute, producing the one or more fashion-ability scores that represents one or more fashion products corresponding to the selected attribute;
- and
- providing one or more recommendations to the user for the one or more fashion products on the website based upon, at least in part, one or more fashion-ability scores representative of the one or more fashion products associated with the user on the website and the one or more fashion-ability scores representative of the one or more fashion products on the website, wherein the one or more recommendations are presented to the user in order of the priority score assigned to the one or more fashion products on the website.
5 more elements
A computer program product comprising a non-transitory computer readable storage medium having a plurality of instructions stored thereon, which, when executed by a processor, cause the processor to perform operations comprising:
- training, according to a process, a plurality of neural networks from a plurality of images, wherein each neural network corresponds to an attribute of the plurality of images, a category of the plurality of images, or a sub-category of the plurality of images, wherein the process is selected and identified to process one or more images of a fashion category and/or a fashion sub-category to generate one or more fashion-ability scores each comprising a numerical representation of a corresponding fashion product defined in a corresponding image of the corresponding fashion product;
- for each trained neural network, scoring each of the plurality of images against each attribute that the neural network is trained for, and generating one or more scored vectors wherein each scored vector corresponds to a particular attribute used to train the neural network;
- joining each of the scored vectors for each of the plurality of images to form a multi-dimensional vector corresponding to a visual representation of each of the plurality of images;
- associating, at the computing device, one or more fashion products on a website with a user accessing the website, wherein each of the one or more fashion products associated with the user on the website is assigned a priority score based upon, at least in part a fashion browsing sequence of the user;
- generating the one or more fashion-ability scores such that the one or more fashion-ability scores are representative of the one or more fashion products on the website, by selecting an attribute and retrieving the scored vector that is trained for the selected attribute;
- in response to retrieving the scored vector trained for the selected attribute, producing the one or more fashion-ability scores that represents one or more fashion products corresponding to the selected attribute;
- and
- providing one or more recommendations to the user for the one or more fashion products on the website based upon, at least in part, one or more fashion-ability scores representative of the one or more fashion products associated with the user on the website and the one or more fashion-ability scores representative of the one or more fashion products on the website, wherein the one or more recommendations are presented to the user in order of the priority score assigned to the one or more fashion products on the website.
5 more elements
A computing system including one or more processors and one or more memories configured to perform operations comprising:
- training, according to a process, a plurality of neural networks from a plurality of images, wherein each neural network corresponds to an attribute of the plurality of images, a category of the plurality of images, or a sub-category of the plurality of images, wherein the process is selected and identified to process one or more images of a fashion category and/or a fashion sub-category to generate one or more fashion-ability scores each comprising a numerical representation of a corresponding fashion product defined in a corresponding image of the corresponding fashion product;
- for each trained neural network, scoring each of the plurality of images against each attribute that the neural network is trained for, and generating one or more scored vectors wherein each scored vector corresponds to a particular attribute used to train the neural network;
- joining each of the scored vectors for each of the plurality of images to form a multi-dimensional vector corresponding to a visual representation of each of the plurality of images;
- associating, at the computing device, one or more fashion products on a website with a user accessing the website, wherein each of the one or more fashion products associated with the user on the website is assigned a priority score based upon, at least in part a fashion browsing sequence of the user;
- generating the one or more fashion-ability scores such that the one or more fashion-ability scores are representative of the one or more fashion products on the website, by selecting an attribute and retrieving the scored vector that is trained for the selected attribute;
- in response to retrieving the scored vector trained for the selected attribute, producing the one or more fashion-ability scores that represents one or more fashion products corresponding to the selected attribute;
- and
- providing one or more recommendations to the user for the one or more fashion products on the website based upon, at least in part, one or more fashion-ability scores representative of the one or more fashion products associated with the user on the website and the one or more fashion-ability scores representative of the one or more fashion products on the website, wherein the one or more recommendations are presented to the user in order of the priority score assigned to the one or more fashion products on the website.
5 more elements
Dependencies are inferred from the claim text.