AI SoC Neural Network Semiconductor for Real-Time TV Content Annotation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile devices and applications lack the capability to interact with neural network semiconductors to process and annotate consumer content in real-time, failing to provide consumers with relevant information during the consumption of content.
Innovation Solution
The implementation of an artificial intelligence System on Chip (AI SOC) that executes a machine learning model on received televised content to identify objects, allowing for the display of identified objects for selection through a mobile application interface, and modifying the display to include overlays with relevant cloud information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If mobile devices are used to deliver consumer content without processing, then device complexity is reduced, but the capability to provide real-time annotation and relevant information is lost
Solution Approach 1:
The system divides the content delivery function into two segments: the mobile device delivers raw consumer content without processing, while a separate neural network semiconductor performs the annotation and information enrichment. This segmentation allows each component to specialize, reducing the complexity burden on the mobile device while ensuring comprehensive information processing occurs elsewhere in the system.
2Productivity
If neural network processing is implemented on mobile devices, then real-time annotation capability is improved, but device complexity and power consumption increase
Solution Approach 1:
The patent introduces an external neural network semiconductor as an intermediary processing unit that handles the computationally intensive annotation tasks. This mediator receives consumer content from the mobile device, performs real-time processing using specialized neural network hardware, and returns annotated content. This approach enables real-time annotation capability while preventing the mobile device itself from becoming overly complex.
3Ease of operation
If consumer content is delivered without processing, then ease of operation is maintained, but the user experience lacks interactive annotation and relevant information
Solution Approach 1:
The system implements self-service by enabling the content delivery system to automatically perform annotation and information enrichment without requiring user intervention. The neural network semiconductor autonomously processes consumer content, identifies objects, scenes, and entities, and overlays relevant information, allowing the system to serve itself and the user simultaneously without adding operational complexity.
Data Source
AI summary
Systems and methods described herein involve executing, using an artificial intelligence System on Chip (AI SoC), a machine learning model on received televised content, the machine learning model configured to identify objects displayed on the received televised content; displaying, through a mobile application interface, the identified objects for selection; and for a selection of one or more objects from the identified objects and an overlay through the mobile application interface, modifying a display of the received televised content to display the overlay.