Gradient-Guided Exploration of Generative Model's Latent Space for Controlled Iris Image Augmentations

2025-11-12

AI Summary

To address the scarcity of high-quality, diverse synthetic data for iris recognition and presentation attack detection, this paper proposes a gradient-guided latent space traversal method. Leveraging pre-trained generative adversarial networks (GANs) and GAN inversion frameworks, it optimizes latent codes via differentiable feature losses to enable precise, controllable editing of geometric and quality attributes—including sharpness, pupil/iris size, and aspect ratios—without model fine-tuning. The approach is compatible with both real and synthetic iris images and preserves identity consistency. It supports arbitrary differentiable attribute loss definitions, ensuring strong generalizability and extensibility. Experiments demonstrate high-fidelity image generation, low attribute control error, and substantial improvements in data diversity and downstream task robustness.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Natural Language Processing: Generation
• Machine Learning: Feature Construction/Reformulation

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance

Abstract

Developing reliable iris recognition and presentation attack detection methods requires diverse datasets that capture realistic variations in iris features and a wide spectrum of anomalies. Because of the rich texture of iris images, which spans a wide range of spatial frequencies, synthesizing same-identity iris images while controlling specific attributes remains challenging. In this work, we introduce a new iris image augmentation strategy by traversing a generative model's latent space toward latent codes that represent same-identity samples but with some desired iris image properties manipulated. The latent space traversal is guided by a gradient of specific geometrical, textural, or quality-related iris image features (e.g., sharpness, pupil size, iris size, or pupil-to-iris ratio) and preserves the identity represented by the image being manipulated. The proposed approach can be easily extended to manipulate any attribute for which a differentiable loss term can be formulated. Additionally, our approach can use either randomly generated images using either a pre-train GAN model or real-world iris images. We can utilize GAN inversion to project any given iris image into the latent space and obtain its corresponding latent code.

Problem

• Synthesizing same-identity iris images with controlled specific attributes
• Developing reliable iris recognition and presentation attack detection methods
• Manipulating iris image properties while preserving identity representation

Innovation

• Gradient-guided latent space traversal for iris augmentation
• Manipulates iris features while preserving identity
• Works with GAN-generated images and real iris images