View source for 18182467. Multiple Dataset Search Based On a Visual Query simplified abstract (Google LLC)

=Multiple Dataset Search Based On a Visual Query=


==Organization Name==

[[:Category:Google LLC|Google LLC]]


[[Category:Google LLC]]

==Inventor(s)==

[[:Category:Utsav Lathia of San Francisco CA (US)|Utsav Lathia of San Francisco CA (US)]][[Category:Utsav Lathia of San Francisco CA (US)]]

[[:Category:Sundeep Vaddadi of Los Gatos CA (US)|Sundeep Vaddadi of Los Gatos CA (US)]][[Category:Sundeep Vaddadi of Los Gatos CA (US)]]

==Multiple Dataset Search Based On a Visual Query - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18182467 titled 'Multiple Dataset Search Based On a Visual Query

The abstract of this patent application describes systems and methods that utilize an embedding model to generate an image embedding for image data, which can then be used to determine relevant search results in multiple datasets. The search can be a pure embedding search for one dataset and a multimodal search for another, based on user and image contexts.

* Leveraging an embedding model to generate image embeddings for image data
* Determining relevant search results in multiple datasets
* Pure embedding search for one dataset and multimodal search for another
* Selection of datasets based on user and image contexts
* Providing search results simultaneously to a user computing system

==Potential Applications==
The technology can be applied in image search engines, e-commerce platforms, and content recommendation systems.

==Problems Solved==
This technology addresses the challenges of accurately retrieving relevant search results from diverse datasets based on image content.

==Benefits==
The benefits of this technology include improved search result relevance, enhanced user experience, and more efficient information retrieval.

==Commercial Applications==
* Image search engines
* E-commerce platforms
* Content recommendation systems

==Questions about Image Embedding Technology==
===How does the embedding model generate image embeddings?===
The embedding model uses a mathematical technique to represent images as vectors in a high-dimensional space, capturing their visual features.

===What are the potential limitations of using image embeddings for search purposes?===
One potential limitation could be the accuracy of the embedding model in capturing all relevant visual information from images.


==Original Abstract Submitted==

Systems and methods disclosed herein can leverage an embedding model to generate an image embedding for image data. The image embedding can then be utilized to determine relevant search results in each of a plurality of datasets. The systems and methods may include a pure embedding search for one dataset and a multimodal search for another dataset. One or more of the datasets may be selected for search based on one or more contexts associated with the user and/or the image. The search results may then be provided simultaneously to a user computing system.



[[Category:G06F16/532]]
[[Category:G06F16/538]]
[[Category:G06F16/583]]
[[Category:G06V10/77]]
[[Category:CPC_G06F16/532]]