Dynamic Image Generation for Context-Aware Ad Displays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing graphical advertisement displays use static images that are not tailored to specific user devices, applications, or publisher contexts, leading to suboptimal ad performance.
Innovation Solution
A computer-implemented method and system that dynamically generates graphical display source code by altering seed images based on user device capabilities and context-specific features to select optimized candidate images for inclusion in advertisements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If static images are used for computerized graphical advertisement displays, then the system complexity is reduced and ease of manufacture is improved, but the adaptability to specific user devices, applications, and publisher contexts deteriorates
Solution Approach 1:
The patent implements dynamic image generation by creating multiple candidate images from seed images through automated alterations, then selecting the optimal image based on real-time context factors such as user device capabilities, application type, and publisher context. This transforms the static image selection process into a dynamic, adaptive system that responds to varying conditions.
Solution Approach 2:
The system alters parameters of seed images (such as size, format, color scheme, or content variations) to generate multiple candidate images. By changing these parameters systematically, the system creates a portfolio of adapted images without requiring manual creation of each variant, thus maintaining ease of manufacture while improving adaptability.
2Adaptability or versatility
If multiple customized candidate images are generated based on context and seed image alterations, then the adaptability and performance optimization are improved, but the device complexity and processing time increase
Solution Approach 1:
The image generation process is segmented into distinct modular steps: retrieving seed images, altering them through various transformations, scoring the resulting candidate images based on context, and selecting the optimal image. This segmentation allows each component to be independently optimized and managed, reducing overall system complexity despite the sophisticated functionality.
Solution Approach 2:
Instead of creating entirely new images from scratch, the system copies and modifies existing seed images to generate candidate images. This approach leverages pre-existing high-quality image assets and reduces the computational complexity of image generation while maintaining adaptability through systematic alterations.
3Reliability
If multiple candidate images are generated and scored based on context, then the ad performance and interaction probabilities are improved, but the processing time and productivity are reduced
Solution Approach 1:
The system generates a limited set of candidate images (not all possible variations) and scores them to identify the top candidates. By performing partial action—generating and evaluating only the most promising candidates rather than exhaustively processing all possible image variations—the system achieves good performance optimization without excessive processing time.
Solution Approach 2:
The system pre-generates multiple candidate images from seed images before the actual ad display decision is needed. This preliminary action allows the scoring and selection process to work with pre-prepared materials, reducing the time required at the moment of ad deployment while maintaining the ability to select optimized images.
Data Source
AI summary
A computer receives a request for graphical display source code for a computerized graphical advertisement display, and retrieves seed images including a plurality of seed image features. The computer generates candidate images based on the one or more seed images, where the computer alters a first aspect of a seed image to generate an altered seed image having a plurality of altered seed image features and the computer alters a second aspect of the altered seed image to generate a candidate image having a plurality of candidate image features. The computer generates candidate image scores based upon a context of the advertisement display and the plurality of candidate image features. The computer selects an image from the candidate images based on the candidate image scores and generates the graphical display source code based on the selected image, a size of the advertisement display, and display capabilities of the user device.