AR Ad Delivery Using Pre-Extracted Recognition Indices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality systems face high processing loads when displaying advertisements in real-world environments by extracting local features from natural video images, making it inefficient to specify and deliver advertisements to precise locations.
Innovation Solution
An augmented reality notification information delivery system that includes a delivery control device and a display terminal, using recognition indices and position information to determine if an advertisement space is within the user's field of view, allowing for efficient transmission and display of advertisements without excessive processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If local feature extraction (SIFT/SURF) is used to display advertisements in real-world video images, then advertisement positioning precision is improved, but processing load on the device increases significantly
Solution Approach 1:
The system pre-extracts local features from background images and stores them in a database before actual advertisement delivery. When a user requests an advertisement, the system retrieves pre-processed feature data instead of extracting features in real-time, significantly reducing the processing load on the user's device while maintaining positioning precision.
Solution Approach 2:
The invention introduces a server as an intermediary between the advertisement delivery system and the user's augmented reality device. The server performs the computationally intensive local feature extraction and matching operations, acting as a mediator that offloads processing from the user's device. The server sends only the necessary advertisement data and pre-processed feature information to the user's device, reducing its processing burden.
2Manufacturing precision
If local feature extraction is performed in real-time to match advertisements with real-world locations, then advertisement delivery accuracy is improved, but processing time increases
Solution Approach 1:
The system performs local feature extraction on background images in advance and stores the extracted features in a database. When delivering advertisements, the system retrieves pre-extracted features and performs only lightweight matching operations, dramatically reducing processing time while maintaining delivery accuracy.
Solution Approach 2:
The system pre-processes and stores local features from background images before actual advertisement delivery occurs. This advance preparation allows the system to quickly match advertisements with locations by comparing against pre-extracted features, rather than extracting features in real-time during advertisement delivery.
Data Source
AI summary
A delivery control device generates a recognition index on a notification space basis, the recognition index used to recognize the notification space, displays the recognition index in the notification space, and stores the recognition index in management information. The delivery control device instead stores an image characteristic extracted from a content to be displayed in the notification space as the recognition index in the management information. In this state, the delivery control device selects a notification space viewable on the augmented reality display terminal based on information on the position of the augmented reality display terminal held by a user and information on the position of the notification space and transmits the recognition index corresponding to the selected notification space and a notification content to the augmented reality display terminal. In response to the above, the augmented reality display terminal extracts the recognition index from the captured real-life video images and superposes the notification content in the position corresponding to the extracted recognition index in the real-life video images to generate and display augmented reality notification video images.