Image Authority Scoring via Resource Quality and Similarity Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image search technologies fail to effectively identify and rank high-quality images of entities, such as individuals, based on their inclusion in multiple high-quality resources and similarity to other images, leading to a lack of authoritative and visually representative results.
Innovation Solution
The method involves generating image resource quality scores from quality scores of resources containing images, identifying similar images, and calculating image scores and authority scores to rank images based on their quality and similarity, with facial recognition confidence and portrait scores contributing to the final ranking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image search results are ranked based on simple relevance metrics, then the search process is fast and simple, but the quality and authority of identified images cannot be effectively determined
Solution Approach 1:
The image quality assessment is segmented into multiple independent scoring components: image resource quality scores (based on source website authority), image scores (based on visual quality metrics), and similarity scores (based on comparison with other images). Each component is calculated separately and then combined to produce a comprehensive image authority score, allowing precise measurement without overwhelming complexity
Solution Approach 2:
The patent introduces intermediary scoring mechanisms that mediate between simple relevance metrics and complex quality assessment. Image resource quality scores serve as intermediaries that capture source authority, while similarity scores mediate between individual image evaluation and overall image set quality. These intermediaries enable progressive refinement of image ranking without requiring all complexity at once
2Reliability
If multiple scoring factors are used to determine image authority, then image quality and authority can be accurately assessed, but the computational complexity and processing time increase
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing image resource quality scores based on the authority of source websites, and pre-identifying similar images for each image in the collection. These preliminary scores and relationships are stored and can be quickly retrieved during search operations, eliminating the need to recalculate them from scratch and significantly reducing processing time while maintaining reliable multi-factor assessment
Solution Approach 2:
The scoring system is designed to be dynamic and adaptive. The image authority score is computed on-demand by combining pre-computed components (image resource quality score, image score, similarity score) in a flexible formula. This allows the system to adjust the weight and combination of scoring factors based on specific search needs while maintaining computational efficiency through the modular structure
Data Source
AI summary
Methods, systems, and apparatus for scoring images related to entities. In one aspect, a method includes identifying images associated with a person, each image being included in one or more resources; obtaining, for each resource that includes one of the images, a quality score that represents a quality of the resource; for each of the images: generating an image resource quality score from the quality scores of the resources that include the image; identifying a set of similar images from the images, each similar image having a measure of similarity to the image that meets a similarity measure threshold; generating an image score based on image resource quality scores of the resources that include the similar images relative to image resource quality scores of the resources that include each of the images; and generating an image authority score based on the image resource quality score and the image score.