Current iris recognition technology faces practical difficulties. For example, due to the unsteady morphology of a heterogeneous iris generated by a variety of different devices and environments, the traditional processing methods of statistical learning or cognitive learning for a single iris source are not effective. The existing iris data set size and situational classification constraints make it difficult to meet the requirements of learning methods under a single deep learning framework. Therefore, this paper proposes a method of heterogeneous iris recognition based on an entropy feature lightweight neural network under multi-source feature fusion. The method is divided into an image-processing module and a recognition module. The image-processing module converts the iris image into a recognition label via a convolutional neural network. The recognition module is based on statistical learning ideas and design of a multi-source feature fusion mechanism. The information entropy of the iris feature label is used to set the iris entropy feature category label and design the recognition function according to the category label to obtain the recognition result. As the requirement for the number and quality of irises changes, the category labels in the recognition function are dynamically adjusted using a feedback learning mechanism. This paper uses iris data collected from three different devices in the JLU iris library. The experimental results prove that for multi-category classification of lightweight constrained multi-state irises, the abovementioned problems are ameliorated to a certain extent by this method.

In this paper, the research focuses on the constrained iris under different acquisition states, and trains the mixed data without manual label division. A multi-state iris multi-classification recognition method based on convolutional neural network fusion statistical cognitive learning is proposed. The recognition process is divided into an image processing module, a classification module and a result output module. The image processing module converts iris images into recognition tags through convolutional neural network. By combining the characteristics of cognitive learning and statistical learning, the classification module converts iris features of the same category to the convolutional neural network label parameters by data statistical, and forms a single iris cognitive concept, thereby designing a single-category recognizer. Multiple single recognizers are combined in parallel to perform multi-classification recognition by the codeless fusion recognition mode, the results are further encrypted by external encryption. The final classification result is exported through the result output module, which can improve the security of data transmission. After confirming stealing attacks, the result output security protocol is activated, and legitimate users who pass the reliable third-party authentication can still get correct results after changing the decryption key and the result output mapping path. In the multi-state iris recognition, the experimental results of JLU iris library demonstrate that the proposed method can ensure the correct rate of the single classifier certification. In multi-classification recognition, it can effectively cluster iris of the same category and distinguish different categories of irises. In addition, the method can also effectively respond to the stealing attack behavior.

Precise iris segmentation is a very important part of accurate iris recognition. Traditional iris segmentation methods require complex prior knowledge and pre- and post-processing and have limited accuracy under non-ideal conditions. Deep learning approaches outperform traditional methods. However, the limitation of a small number of labeled datasets degrades their performance drastically because of the difficulty in collecting and labeling irises. Furthermore, previous approaches ignore the large distribution gap within the non-ideal iris dataset due to illumination, motion blur, squinting eyes, etc. To address these issues, we propose a three-stage training strategy. Firstly, supervised contrastive pretraining is proposed to increase intra-class compactness and inter-class separability to obtain a good pixel classifier under a limited amount of data. Secondly, the entire network is fine-tuned using cross-entropy loss. Thirdly, an intra-dataset adversarial adaptation is proposed, which reduces the intra-dataset gap in the non-ideal situation by aligning the distribution of the hard and easy samples at the pixel class level. Our experiments show that our method improved the segmentation performance and achieved the following encouraging results: 0.44%, 1.03%, 0.66%, 0.41%, and 0.37% in the Nice1 and 96.66%, 98.72%, 93.21%, 94.28%, and 97.41% in the F1 for UBIRIS.V2, IITD, MICHE-I, CASIA-D, and CASIA-T.

The demand for applying the iris segmentation model on mobile devices has been growing rapidly. Most current segmentation networks have an enormous amount of parameters, hence unsuitable for mobile devices, while other small memory footprint models follow the spirit of classification networks and ignore the inherent characteristic of segmentation. To address the challenge, we propose a lightweight segmentation network (LiSeNet) for iris segmentation of noisy images. Unlike previous studies that only focus on improving the accuracy of segmentation masks, LiSeNet can simultaneously obtain segmentation masks, parameterized pupillary and limbic boundaries of the iris, further enabling CNN-based iris segmentation to be applied in any regular iris recognition systems. We first propose a multiscale concatenate (MSC) Block, which connects multiple sizes of convolution kernels in a dense manner, gradually reduces the dimension of feature maps and uses the aggregation of them for image representation. Based on the MSC block, we develop a two-stage refinement encoder to aggregate discriminative features through subnetwork feature reuse and substage feature reassess, thus obtaining a sufficient receptive field and enhancing the model learning ability. To exploit object contextual information more efficiently, we further devise a grouped spatial attention to emphasize the important features and suppress irrelevant noises through a gating mechanism in the decoder. Extensive experiments on three challenging iris datasets show that LiSeNet, without any complicated postprocessing, achieves competitive or state-of-the-art performance with only 2.2M parameters, being 14 × smaller than the previous best method. Code will be publicly available.