Product-Inclusive Image Alt Text from Brand and Product Inputs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual and software-based image description methods are costly, time-consuming, and often inaccurate, especially for complex images with non-simple backgrounds.
Innovation Solution
A system utilizing a multimodal generative artificial intelligence (GenAI) model to generate accurate and comprehensive image alt text by extracting brand and product information from user-submitted inputs, and optionally incorporating promotional details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual creation of text descriptions is used, then accuracy of image description is improved, but cost and time consumption increase
Solution Approach 1:
The patent introduces an intermediary system comprising image processing algorithms and natural language generation models that act as a mediator between the input image and the final text description. This intermediary processing layer automatically extracts visual features, identifies objects and attributes, and generates descriptive text, thereby eliminating the need for manual description creation while maintaining high accuracy through multi-stage processing and validation mechanisms.
2Productivity
If software-based image description generation is used, then time consumption is reduced, but accuracy deteriorates especially for complex images
Solution Approach 1:
The patent segments the image description generation process into multiple independent modules: image pre-processing module, feature extraction module, object identification module, attribute recognition module, and text generation module. Each module specializes in specific tasks and can be independently optimized. This segmentation allows the system to handle complex images more accurately by breaking down the comprehensive analysis into manageable specialized steps, thereby improving both efficiency and accuracy simultaneously.
Solution Approach 2:
The patent implements dynamic adjustment mechanisms where the processing depth and algorithm selection adapt based on image complexity. For simple images, the system uses streamlined processing paths for speed, while for complex images, it automatically activates more sophisticated analysis routines including multi-scale feature extraction and contextual reasoning. This dynamic adaptability ensures high accuracy for complex images while maintaining efficiency for simpler cases.
3Speed
If simple image processing algorithms are used, then processing speed is improved, but description completeness deteriorates
Solution Approach 1:
The patent performs preliminary actions by pre-processing images to enhance key features before main processing, and by pre-training models on extensive datasets to capture common patterns. The system also performs preliminary segmentation and feature extraction that prepares the data for subsequent detailed analysis. These preliminary actions enable the system to achieve comprehensive descriptions more efficiently by reducing the computational burden during the main processing stages.
Data Source
AI summary
A method can be implemented via execution of computing instructions configured to run at a processor. The method can include: receiving, from a user, an image of a product; receiving, from the user, user-submitted logo alt text describing a brand of the product in the image; receiving, from the user, user-submitted image alt text describing the image; extracting brand information from the user-submitted logo alt text; extracting product information from the user-submitted image alt text; generating an instruction prompt that includes the brand information, as extracted, and the product information, as extracted; and generating a recommended image alt text describing the image and including the brand information, as extracted, and the product information, as extracted, by querying a multimodal generative artificial intelligence (multimodal GenAI) model with the instruction prompt. Other embodiments are disclosed.