OcclusionNetPlusPlus: a multi-scale similarity network with adaptive occlusion detection for robust iris recognition

Innovation in Architecture We've developed a uniqueframework that combines multi-scale Gabor filters with an intelligent occlusion detection system. Unlike traditional methods that simply mask out occluded regions, our approach uses probability maps to adaptively weight features based on occlusion likelihood. This preserves valuable information from partially occluded regions while minimizing their potential negative impact. Technical Innovations :- Occlusion-aware iris recognition framework combining multi-scale Gabor filters with dynamic probability maps. Intelligent occlusion detection and adaptive feature weighting that preserves information from partially occluded regions.Attention pooling and a custom loss for improved feature selection and robust similarity computation.Validated on CASIA-Iris-Thousand and IITD (≈22k images) with ROC AUC = 0.9948 and D-prime = 6.433.Suitable for high-security deployments; extendable to lightweight mobile variants and anti-spoofing enhancements. Real-World Applications The framework is particularly suited for: • High-security access control systems • Border security and immigration control • Financial transaction authentication • Healthcare facility security systems • Corporate access management Impact and Significance Our work represents a significant advancement in making iris recognition more practical and reliable for real-world applications. The framework's ability to handle occlusions while maintaining high accuracy addresses one of the major challenges in biometric security systems. This breakthrough has important implications for both current deployments and future developments in biometric authentication. Future Directions While achieving significant advances, our research opens exciting possibilities for: • Extension to completely unconstrained environments • Integration with other biometric modalities • Development of lightweight variants for mobile applications • Enhancement of anti-spoofing capabilities This research significantly advances the field of iris recognition, particularly in handling real-world occlusion scenarios while maintaining exceptional accuracy and reliability. Our comprehensive approach combines sophisticated occlusion detection with adaptive feature processing, setting a new standard for biometric security systems. A significant challenge in iris recognition systems is the presence of occlusions affecting the iris, face, and periocular regions. To address this issue, this study proposes an OcclusionNetPlusPlus framework which employs carefully designed bank of Gabor filters to capture iris texture patterns at different scales and orientations. We then inject 2D positional encodings into these filter responses to embed explicit (x,y) location information, enabling downstream modules to reason about where each feature came from. The innovation in our approach is the introduction of an occlusion detection mechanism that generates probability maps based on local variance analysis, effectively identifying occluded regions in the iris image. These probability maps are used to dynamically weight the extracted features, reducing the influence of unreliable regions during similarity computation. The framework incorporates a custom loss function that optimizes feature similarity while maintaining discriminative power across different iris patterns. Training and evaluation were conducted on publicly available iris recognition datasets, ensuring a diverse test bed for assessing performance across different occlusion scenarios. We evaluated OcclusionNetPlusPlus on CASIA-Iris-Thousand and IIT Delhi V1.0. In controlled tests, it achieves an EER of 0.51 %, an FRR of 0.54 % at FAR = 1 % (0.61 % at FAR = 0.1 %), and a d-prime of 7.04. Even under simulated unconstrained conditions—adding noise, blur, and random occlusions—EER stays around 2 %.