Evaluating Conformance of Video Safety Tools for Photosensitive Epilepsy.

Area of Science:

• Neurology and Digital Media Safety
• Epilepsy Research
• Human-Computer Interaction

Background:

• Photosensitive epilepsy (PSE) is a neurological condition where flashing lights or specific visual patterns can trigger seizures.
• International guidelines exist to define harmful visual stimuli, focusing on area, intensity, and flash rate.
• Automated tools are used to assess video content for PSE risks, but lack sufficient validation.

Purpose of the Study:

• To identify limitations in current testing methodologies for photosensitive epilepsy (PSE) hazard analysis tools.
• To propose enhanced validation tests for PSE assessment algorithms.
• To introduce a novel tool for generating ground-truth video sequences for PSE algorithm validation.

Main Methods:

• Analysis of existing testing paradigms for automated PSE risk assessment tools.
• Development of a new tool capable of generating diverse, ground-truth video sequences.
• Creation of test videos that go beyond simple alternating-frames, including complex patterns and confounding stimuli.

Main Results:

• Current validation methods and test videos for PSE analysis tools are insufficient.
• The developed tool can generate advanced video sequences for more rigorous algorithm testing.
• The new test videos address gaps in evaluating adherence to international PSE guidelines.

Conclusions:

• There is a critical need for improved validation strategies and resources for PSE safety tools.
• The proposed validation tests and the new video generation tool enhance the reliability of PSE hazard analysis.
• This work contributes to safer digital media consumption for individuals with photosensitive epilepsy.