Amazon Technologies, Inc.

Zohar Barzelay

Nikhil Garg

Brad Bowman

Michael Donoser

Sabine Sternig

Toma Belenzada

Bojan Pepik

Matthieu Guillaumin

Various embodiments of systems and methods allow a system to identify subsets of items by mixing and matching identified features in one or more other items. A system can identify features of items in an item database. The system can then calculate 'fingerprints' of these features which are vectors describing the characteristics of the features. The system can present a collection of items and a user can select an item of the collection. The user can then select positive features to include in a search and/or negative features to include in the search. The system can then do a search of the database for items that contain features similar to those positive features and do not contain features similar to those negative features. The user can select features through a variety of means.

Feature-based search

Techniques are generally described for predicting text relevant to image data. In various examples, the techniques may include receiving image data comprising a first portion. The first portion of the image data may correspond to a first plurality of pixels when rendered on the display. Text data comprising a first text related to the first portion of the image data may be received. A first vector representation of the first portion of the image data may be determined. In some examples, a correspondence between the first portion of the image data and the first text may be determined based at least in part on the first vector representation. A first identifier of the first portion of image data may be stored in a data structure in association with a second identifier of the first text.

Relevant text identification based on image feature selection

Techniques for providing an in-place scrolling feature are described herein. Input via a user interface may be received to update first images of first items to second images of second items where the first images are presented in a focus location of the user interface. Feature vectors may be calculated for each image associated with the first items and the second items using a neural network algorithm. The feature vectors may include a plurality of values that represent each image of the first items and the second items. Candidate images may be determined from the second images based on a distance in a feature vector space between the feature vectors of the candidate images and the feature vectors of the second images. The user interface may be updated to transition from presenting the first images to a portion of the candidate images in the focus location of the user interface.

In-place scrolling for a user in interface

Techniques for prioritizing images associated with an item to display an appropriate image based on a query are described herein. For example, an attention score for an item attribute in an image of an item may be generated based at least in part on a model that uses one or more images of the item. The item attribute for the item associated with a query may be obtained. A plurality of items may be determined based at least in part on the item attribute being associated with the plurality of items where an individual item of the plurality of items includes a plurality of images of the item. The plurality of images of the individual item may be ranked based at least in part on corresponding attention scores associated with each image of the plurality of images.

Growing community of inventors

Potsdam, Germany

Zohar Barzelay