Media Asset Image Personalization via Entity Prevalence Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Media guidance applications lack the ability to personalize images associated with media assets based on user preferences, leading to non-relevant content being displayed.
Innovation Solution
Control circuitry identifies user preferences from a profile and uses image recognition algorithms to crop or generate images that include only preferred entities from a media asset's poster art, ensuring that images displayed are more relevant to the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If generic poster art is displayed for media assets, then all entities in the image are shown, but the image is not personalized to user preferences
Solution Approach 1:
The system segments the poster art image into multiple portions, each containing a different entity (actor, character, or logo). Image recognition algorithms analyze the image to identify and separate these entities, allowing the system to selectively display only the portions corresponding to entities the user prefers, thus achieving personalization while preserving relevant information.
2Loss of information
If the entire poster art image is displayed, then all entities are included, but non-preferred entities reduce relevance to the user
Solution Approach 1:
The system extracts and removes portions of the poster art that correspond to entities the user does not prefer. By using image recognition to identify entity locations and then cropping out the unwanted portions, the system displays only the relevant entities, eliminating information loss while maximizing relevance to user preferences.
3Adaptability or versatility
If image cropping is performed to personalize display, then user preference alignment improves, but image processing complexity increases
Solution Approach 1:
Instead of permanently modifying the original poster art image, the system creates cropped copies or thumbnails that contain only the preferred entities. This approach allows the system to achieve accurate personalization while keeping the processing relatively simple, as the cropping operations are performed on duplicate image data rather than the master file.
Data Source
AI summary
Control circuitry may identify an image corresponding to a media asset, where the image comprises a plurality of portions, and where each portion of the plurality of portions comprises a respective entity. The image identified by control circuitry may include several portions, where each portion includes a respective entity. Control circuitry may determine an identity of each respective entity in each respective portion of the plurality of portions. Control circuitry may then access a database to determine a level of prevalence of each identity in the media asset. Based on these identifications, control circuitry may determine which respective entity corresponds to a highest level of prevalence (e.g., with respect to the other entities), and may generate for display a respective portion comprising the respective entity that corresponds to the highest level of prevalence.