US 12,425,694 B2
Enabling a more accurate search of a digital media database
Harshith Bariki, Bangalore (IN); Siddarth Tirumakudlu Vinay, Bangalore (IN); and Joel Vivek Iniyan R, Bangalore (IN)
Assigned to DISH Network Technologies India Private Limited, Bangalore (IN)
Filed by DISH Network Technologies India Private Limited, Bangalore (IN)
Filed on Dec. 29, 2023, as Appl. No. 18/401,246.
Prior Publication US 2025/0220278 A1, Jul. 3, 2025
Int. Cl. H04N 21/482 (2011.01); G06F 16/783 (2019.01); G06F 40/40 (2020.01); H04N 21/472 (2011.01); H04N 21/81 (2011.01); H04N 21/854 (2011.01)
CPC H04N 21/4828 (2013.01) [G06F 16/7844 (2019.01); G06F 40/40 (2020.01)] 20 Claims
OG exemplary drawing
 
1. A non-transitory, computer-readable storage medium comprising instructions recorded thereon, wherein the instructions, when executed by at least one data processor of a system, cause the system to:
obtain, from a database storing multiple video data, a video data among the multiple video data including metadata associated with the video data,
wherein the video data is associated with a video,
wherein the metadata includes a title associated with the video data,
wherein the database storing the multiple video data is configured to support a first search using the metadata;
extract, from the video associated with the video data, an audio and a closed caption data;
provide the audio, the closed caption data, the title associated with the video data, and a prompt to a large language model,
wherein the prompt requests multiple tags based on the audio, the closed caption data, and the title associated with the video data,
wherein a tag among the multiple tags includes a natural language text indicating content associated with the video data;
obtain the multiple tags from the large language model;
determine relevance associated with the multiple tags;
based on the relevance, select a predetermined number of tags from the multiple tags;
decrease a memory footprint associated with the database by storing the selected tags in the database by adding the selected tags to the metadata associated with the video data to obtain new metadata; and
enable a second search of the multiple video data stored in the database by searching the new metadata,
wherein results of the second search have a lower error margin than results of the first search.