Parallel scheme for real-time detection of photosensitive seizures

Area of Science:

• Computer Science
• Neurology
• Digital Media

Background:

• The rise of online video content creation and distribution poses health risks.
• Individuals with photosensitive epilepsy are vulnerable to seizures triggered by specific visual stimuli.
• Existing methods for detecting hazardous video content are insufficient for real-time application.

Purpose of the Study:

• To develop and evaluate a novel spatiotemporal pattern detection algorithm.
• To enable real-time detection of seizure-inducing content in streaming videos.
• To compare the proposed algorithm's performance against existing post-processing tools.

Main Methods:

• Development of a spatiotemporal pattern detection algorithm for real-time analysis.
• Creation of a tool for generating test videos with hazardous content.
• Implementation of the algorithm on a dual-core processor using a pipelined/parallel architecture.
• Evaluation using generated test videos and comparison with a current post-processing tool.

Main Results:

• The proposed algorithm demonstrates superior detection performance compared to the existing tool.
• Real-time detection of seizure-inducing patterns is achieved.
• The algorithm effectively masks hazardous visual patterns.

Conclusions:

• The developed algorithm offers an effective solution for real-time detection of hazardous video content.
• This technology can significantly improve the safety of online video consumption for individuals with photosensitive epilepsy.
• Further optimization for real-time streaming applications is feasible.