Multimodal Alt-Text Evaluation for Image Accessibility Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Web pages often provide poor-quality alt-text that does not accurately describe images, leading to a poor user experience for blind and visually impaired users and negatively affecting search engine optimization (SEO) performance.
Innovation Solution
A system using a multi-modal text/image model, such as the CLIP model, evaluates alt-texts by calculating similarity and complexity scores to select the best alt-text that accurately describes images, enhancing accessibility and SEO.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual alt-text writing is used, then implementation is simple, but quality and accuracy are poor
Solution Approach 1:
The patent introduces an automated evaluation system as an intermediary between image content and alt-text generation. This system uses machine learning models to objectively assess alt-text quality, serving as a mediator that guides improvement without requiring direct manual intervention in the writing process itself.
Solution Approach 2:
The patent implements a feedback mechanism where the evaluation system provides quality assessments of existing alt-text, enabling iterative improvement. The system calculates quality scores and provides actionable feedback to help improve alt-text accuracy and relevance automatically.
2Reliability
If automated evaluation system is implemented, then alt-text quality improves, but system complexity increases
Solution Approach 1:
The patent enables the evaluation system to assess and improve its own output quality through automated feedback loops. The system can self-evaluate generated alt-text and iteratively improve without requiring constant external validation, reducing the need for complex external quality assurance mechanisms.
3Productivity
If poor-quality alt-text is used, then implementation is fast, but user experience deteriorates
Solution Approach 1:
The patent applies preliminary action by evaluating and improving alt-text quality before deployment. The automated evaluation system assesses potential alt-text options in advance, ensuring high quality is achieved prior to implementation, thus maintaining both speed and quality through proactive rather than reactive quality control.
4Loss of time
If low-quality alt-text is provided, then processing time is reduced, but search engine performance suffers
Solution Approach 1:
The patent changes the quality parameters of alt-text through automated evaluation and optimization. By adjusting parameters such as accuracy, relevance, and descriptive quality through machine learning-guided improvements, the system achieves better SEO performance without proportionally increasing processing time.
Data Source
AI summary
One example method includes a machine-learning (ML) model receiving a first input that includes images that have been extracted from a web page and a second input that includes alt-texts that have been extracted from the web page. The alt-texts describe the images. The ML model converts the images into a first embedding representation and converts the alt-texts into a second embedding representation. Based on the first and second embedding representations, a similarity score between the images and the alt-texts is calculated. The similarity score specifies how accurately each of the alt-texts describe the images. The one of the alt-texts having the highest similarity score is then selected.