US 2026/0289622 A1 · Image Duplicate Detection Using a Learning Model

Abstract

Image duplicate detection using a learning model is described. A computing device obtains multiple images. The computing device generates respective representations of the images based on providing the images as input to a learning model, which can be stored at the local data storage of the computing device. In some examples, the computing device receives one or more parameters representative of the learning model from a server device. The computing device outputs an indication of one or more images based on the respective representations of the images matching at least one representation of the representations of the other images within a threshold value. The representations of the other images are stored at the local data storage of the computing device.

Claims20

Dependent claims

A computer-implemented method comprising:

obtaining, at a computing device, a first plurality of images;
generating, based at least in part on providing the first plurality of images as input to a learning model, respective representations of the first plurality of images;
and
outputting, at the computing device, an indication of one or more images based at least in part on the respective representations of the one or more images of the first plurality of images matching at least one representation of a plurality of representations of a second plurality of images within a threshold value, wherein the plurality of representations of the second plurality of images are stored at a local data storage of the computing device.

1 more element

A computer-implemented method comprising:

transmitting, to a computing device for storage at a local data storage of the computing device, one or more parameters representative of a learning model, wherein the learning model is configured to generate respective representations of a first plurality of images;
and
transmitting, to the computing device, instructions to cause the computing device to output an indication of one or more images of the first plurality of images based at least in part on the respective representations of the one or more images of the first plurality of images matching at least one representation of a plurality of representations of a second plurality of images within a threshold value.

Dependencies are inferred from the claim text.

Image Duplicate Detection Using a Learning Model

Life of this patent

Image Duplicate Detection Using a Learning Model

Life of this patent

Abstract

Claims20