US 12,169,518 B2
Document retrieval using intra-image relationships
Cristopher Flagg, Annandale, VA (US); and Ophir Frieder, Chevy Chase, MD (US)
Assigned to Georgetown University, Washington, DC (US)
Filed by Georgetown University, Washington, DC (US)
Filed on Apr. 15, 2022, as Appl. No. 17/722,156.
Claims priority of provisional application 63/175,972, filed on Apr. 16, 2021.
Prior Publication US 2022/0342928 A1, Oct. 27, 2022
Int. Cl. G06F 16/532 (2019.01); G06F 16/538 (2019.01); G06F 16/56 (2019.01); G06F 16/583 (2019.01); G06V 10/26 (2022.01); G06V 10/74 (2022.01); G06V 10/75 (2022.01); G06V 10/771 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)
CPC G06F 16/532 (2019.01) [G06F 16/538 (2019.01); G06F 16/56 (2019.01); G06F 16/583 (2019.01); G06V 10/267 (2022.01); G06V 10/75 (2022.01); G06V 10/761 (2022.01); G06V 10/771 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)] 20 Claims
OG exemplary drawing
 
5. A method, performed by one or more computing devices, the method comprising:
obtaining a query image representation comprising two or more segmentations of a query image;
generating one or more latent space representations for each of the two or more segmentations of the query image representation;
generating first one or more feature vectors for the first segmentation and second one or more feature vectors for the second segmentation based on the one or more latent space representations;
determining that the first segmentation is a dominant segmentation in the query image representation and the second segmentation is a non-dominant segmentation in the query image representation;
assigning a greater weight to the first one or more feature vectors and a lower weight to the second one or more feature vectors to generate a weighted set of feature vectors;
obtaining a set of external search results which comprises a set of image representations;
generating one or more latent space representations for each image representation in the set of image representations in the set of external search results;
generating third one or more feature vectors based on the one or more latent space representations generated for each image representation in the set of image representations in the set of external search results; and
modifying an order of the set of external search results based on a similarity comparison of the weighted set of feature vectors and the third one or more feature vectors based on the one or more latent space representations generated for each image representation in the set of image representations in the set of external search results.