METHOD AND ELECTRONIC DEVICE FOR DETECTING AI GENERATED CONTENT IN A VIDEO

Smart overview of the Invention

A method for identifying artificial intelligence (AI) generated content in videos involves analyzing a series of frames. The process starts with receiving the video and detecting objects, people, and backgrounds within each frame. By examining pixel-motion information, the method establishes relationships among the detected elements and their corresponding motion data.

Intrinsic Properties Analysis

For each frame, the method identifies intrinsic properties of the objects, people, and backgrounds based on their relationships with pixel-motion information. This analysis is crucial for understanding how these elements should behave in a natural setting, providing a baseline for comparison.

Detection of Inconsistent Motion

The core of the detection process lies in identifying any inconsistent motion within at least one frame. By leveraging the previously established intrinsic properties, the method can pinpoint discrepancies that suggest the presence of AI-generated content, indicating that certain movements do not align with expected behavior.

Electronic Device Implementation

An electronic device designed for this detection method includes memory and processors that execute specific instructions. These components work together to receive video data, analyze frames for objects and motion, and identify inconsistencies that may signal AI-generated content.

Significance of the Technology

This technology addresses the growing need to differentiate between authentic and AI-generated videos, especially as synthetic content becomes increasingly prevalent. By effectively detecting inconsistencies in motion and relationships within video frames, it enhances video integrity verification across various applications.