Agentic data science with the Data Agent Kit
cloud.google.com·7h ago
AI Research
arXiv CS.AI·Aug 11·5 min read
Detecting Clear Contact Lenses for Iris Recognition: A Two-Stage Mask-Guided Attention Approach
TL;DR
A two-stage framework for detecting clear contact lenses in iris recognition was developed, achieving 90.0%-98.8% accuracy and significantly improving verification performance by reducing error rates.
✦ Why It Matters
Implement this detection framework to enhance the accuracy of your iris recognition systems today.
Key Takeaways
- Integrate the two-stage detection framework into existing iris recognition systems.
- Utilize the ConvNeXt-Base model for improved clear lens detection.
- Apply z-score calibration to enhance match score reliability.
- Evaluate the impact of clear lenses on your current verification processes.
How It Works
The proposed framework consists of two stages: the first stage uses a PAD model to detect patterned lenses, while the second stage employs a ConvNeXt-Base model with Mask-Guided Spatial Attention (MGSA). The MGSA module focuses on subtle limbal cues, enhancing the model's ability to differentiate between clear lenses and no lenses.
This dual approach allows for more accurate iris recognition in the presence of clear contact lenses.