Computer Vision Content Targeting via Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users of communications networks often receive irrelevant or unwanted content items, leading to a poor user experience and inefficient use of network resources, as strong connections may also provide content of little interest.
Innovation Solution
The system uses computer vision processing to extract features from images and determine relevant content items, selecting accounts that have previously searched for related descriptors to provide targeted content, thereby reducing unnecessary content delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If content items are provided to all connected accounts, then network activity and content delivery are maintained, but user experience deteriorates due to irrelevant content and network resources are wasted
Solution Approach 1:
The system performs preliminary actions by extracting features from images and determining descriptors before content delivery. Accounts are pre-screened based on their search history for relevant descriptors, so that only interested accounts receive the content. This preliminary filtering prevents wasteful network resource consumption on irrelevant content delivery while ensuring relevant content reaches the right users.
2Ease of operation
If content filtering is applied based on user search history, then relevant content delivery is improved, but system complexity increases due to feature extraction and descriptor matching
Solution Approach 1:
The system extracts key features from images and separates them into distinct descriptors that can be independently matched against user search histories. This extraction approach simplifies the matching process by breaking down complex image data into manageable feature components, making the filtering system more efficient despite the added processing step.
3Measurement precision
If computer vision processing is used to extract features from images, then content accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system extracts only the most relevant features from images rather than processing the entire image data. By identifying and extracting key descriptors that are most likely to match user interests, the system achieves high content matching accuracy while significantly reducing the computational burden and processing time compared to analyzing complete image datasets.
Data Source
AI summary
In certain embodiments, one or more images of an object may be received from a device associated with a first account on a communications network. Features of the object may be extracted based on the one or more images, and one or more content items related to the object may be determined based on the features. A hashtag associated with at least one of the features may be determined. A second account connected to the first account may be selected where the second account previously performed a search for the hashtag on the communications network, and at least one of the one or more content items may be provided to the second account.