| CPC G06F 16/583 (2019.01) [G06F 16/55 (2019.01); G06V 10/44 (2022.01); G06V 10/761 (2022.01)] | 11 Claims |
|
1. A method for recommending an object, comprising:
recognizing, by using a trained neural network, a retrieval feature of a retrieval object from a retrieval image including the retrieval object, wherein the retrieval image is uploaded by a target user through a client device, the trained neural network is trained using a plurality of classified images, and the retrieval feature of the retrieved object is a type of the retrieval object;
obtaining, based on the retrieval feature, at least one retrieval feature image from a first database including a plurality of feature images, wherein each feature image of the plurality of feature images in the first database corresponds to one classification feature of a plurality of classification features, wherein the classification feature represents a classification of each feature image, wherein the retrieval feature includes a first classification feature corresponding to the retrieval object in the plurality of classification features, and wherein the obtaining at least one retrieval image from the first database including the plurality of feature images comprises:
obtaining at least one first feature image corresponding to the first classification feature in the plurality of feature images; and
obtaining the at least one retrieval feature image from the at least one first feature image, wherein the obtaining the at least one retrieval feature image from the at least one first feature image comprises:
obtaining a first similarity between each first feature image of the at least one first feature image and the retrieval image; and
obtaining the at least one retrieval feature image, wherein the first similarity corresponding to each retrieval feature image of the at least one retrieval feature image is greater than a first threshold;
obtaining, based on the at least one retrieval feature image, a target object image set from a second database including a plurality of object images, wherein the obtaining the target object image set from the second database including the plurality of object images comprises:
obtaining a second similarity between each retrieval feature image of the at least one retrieval feature image and each object image of the plurality of object images;
obtaining one or more first object images from the plurality of object images, wherein a maximum value of at least one second similarity corresponding to each first object image of the one or more first object images is greater than a second threshold; and
obtaining the target object image set based on the one or more first object images; and providing the target object image set as one or more recommendation objects to the client device.
|