Deep Learning-Based Feature Extraction in Iris Recognition: Use Existing Models, Fine-tune or Train From Scratch?
This work addresses a practical optimization problem for researchers and practitioners in biometric security, but it is incremental as it compares standard training strategies without introducing a new method.
The study tackled the problem of choosing a training strategy for deep learning-based feature extraction in iris recognition, finding that fine-tuning an existing ResNet-50 model on a large iris dataset yields greater accuracy than using off-the-shelf weights or training from scratch, with the fine-tuned model showing improved performance over previous work.
Modern deep learning techniques can be employed to generate effective feature extractors for the task of iris recognition. The question arises: should we train such structures from scratch on a relatively large iris image dataset, or it is better to fine-tune the existing models to adapt them to a new domain? In this work we explore five different sets of weights for the popular ResNet-50 architecture to find out whether iris-specific feature extractors perform better than models trained for non-iris tasks. Features are extracted from each convolutional layer and the classification accuracy achieved by a Support Vector Machine is measured on a dataset that is disjoint from the samples used in training of the ResNet-50 model. We show that the optimal training strategy is to fine-tune an off-the-shelf set of weights to the iris recognition domain. This approach results in greater accuracy than both off-the-shelf weights and a model trained from scratch. The winning, fine-tuned approach also shows an increase in performance when compared to previous work, in which only off-the-shelf (not fine-tuned) models were used in iris feature extraction. We make the best-performing ResNet-50 model, fine-tuned with more than 360,000 iris images, publicly available along with this paper.