一种新颖的虹膜图像噪声检测方法*

为提高虹膜识别的正确率,针对虹膜图像中存在着眼睫毛和眼睑这两类较难检测的遮挡噪声,在分析现有检测虹膜噪声算法的优缺点后,提出了一套新颖的虹膜图像噪声检测方法：基于Gabor滤波变换的灰度均值法检测睫毛和利用最小二乘法检测眼睑。实验表明,该算法能有效地检测两种遮挡噪声,准确率分别达到95.10%和96.51%,且等错率（EER）指标与已有算法相比最优,提高了虹膜识别系统的整体性能。     

基于彩色分割的虹膜检测

提出了一种基于彩色分割的虹膜检测方法,对于一幅经过准确定位的眼睛窗口图像,它的彩色信息比灰度信息更有助于进行虹膜检测。首先利用图像颜色的饱和度信息将眼睛图像中的眼睛区域与皮肤区域分离,然后利用亮度信息将眼睛区域的眼白和虹膜分离得到虹膜区域,再通过Hough变换进行虹膜检测。结果证明该方法不仅检测率高,而且能适应一定人脸姿势、眼睛视线方向的变化,甚至在眼睛区域较暗的情况下,也能很好地检测到虹膜。     

一种有效的虹膜定位及睫毛检测方法

提出一种有效的虹膜定位及睫毛检测方法。通过把眼睛图像中分割成小的矩形区域,利用找到的这些矩形区域像素平均最小值把眼睛图像进行二值化,找到虹膜区域的内边界;以瞳孔的质心为参考点,在其左右的扇形区域内分别使用修改后的Daugman的检测算子,找到像素值变换大的位置,进而定位出虹膜的外界;使用Gobor滤波器和窗口移动法对睫毛进行有效的检测。通过对大量虹膜图像的实验表明,该方法取得了非常好的结果。     

活体虹膜图像的定位与分割

介绍了一种活体虹膜的定位与分割算法。算法主要分为两部分：圆环的定位与非虹膜区域的去除。本算法根据眼睛的生理特点和数字虹膜图像的实际情况,利用传统定位方法与数学形态学相结合对虹膜区域进行快速而准确的定位,并分别提出了去除眼睑、睫毛和光斑影响的解决方案。算法中也考虑到实际应用可能遇到的影响虹膜定位与分割的问题。实验表明,该算法取得较好的分割结果,并且具有鲁棒性。     

基于SCM的虹膜干扰区域去除算法

为了提高虹膜识别率,提出了一种新的睫毛及眼睑区域定位算法。为了提取呈不同角度分布的睫毛,结合SCM模型和动态交叉熵准则对归一化虹膜图像进行多次迭代,根据图像的灰度分布特性,选取合适迭代结果,对多幅迭代图像进行边缘像素点跟踪融合来获得理想的干扰区域轮廓定位。采用高斯金字塔尺度变换与霍夫变换相结合的方法对眼睑的类椭圆区域进行拟合,进而获得连续的眼睑边缘,实现对归一化虹膜图像中干扰区域的准确定位。实验结果验证了方法的有效性。     

虹膜分割算法评价基准

虹膜识别是生物特征识别中最稳定和最可靠的身份识别方法之一.在虹膜识别的整个流程中,虹膜分割处于预处理阶段,因此虹膜分割结果的好坏将直接影响虹膜识别的精度.自从1993年Daugman第1次提出高性能的虹膜识别系统以来,各种各样的虹膜分割算法陆续提出,尤其是近年来基于深度学习的虹膜分割算法极大地提升了虹膜分割的精度.然而,由于缺乏统一的数据库和评价指标,各种算法的性能比较杂乱而不公平,因此提出了一个公开的虹膜分割评价基准.首先,介绍了虹膜分割的定义和面临的挑战;其次全面梳理了3个有代表性的公开虹膜分割数据库,总结了其特点和挑战性;紧接着定义了虹膜分割的评价指标;然后对传统的和基于深度学习的虹膜分割算法进行了总结,并通过详细的实验对各类算法进行了比较和分析.实验结果表明：当前基于深度学习的虹膜分割算法在准确性上超越了传统的方法.最后,对基于深度学习的虹膜分割算法存在的问题进行了思考和讨论.     

基于改进Canny边缘检测的堆叠货箱分割定位方法

为了提高集装箱卸货的自动化水平,针对集装箱内堆叠货箱难分割定位的问题,提出了一种基于改进Canny边缘检测的堆叠货箱分割定位方法。通过阈值分割和形态学处理进行图像预处理,去除背景干扰,提取堆叠货箱区域图像,基于改进Canny算法对堆叠货箱进行边缘检测,根据堆叠货箱边缘特征进行筛选并基于最小二乘法进行直线拟合,解决边缘线条不连续和虚假边缘问题,对边缘进行区域化处理,以此将堆叠货箱分割成独立的货箱区域,提取每个独立货箱的最小外接矩形,得到货箱中心点的位置信息。实验结果表明,该方法对堆叠货箱有很好的分割效果,定位精度小于5 mm,满足定位精度要求。     

结合Transformer与对称型编解码器的噪声虹膜图像分割方法EI北大核心CSCD

针对少约束场景下采集的虹膜图像容易受到镜面反射、睫毛和头发遮挡、运动和离焦模糊等噪声的干扰,导致难以准确地分割有效的虹膜区域的问题,提出一种结合Transformer与对称型编解码器的噪声虹膜图像分割方法.首先,使用Swin Transformer作为编码器,将输入图像的区块序列送入分层Transformer模块中,通过自注意力机制建模像素间的长距离依赖,增强上下文信息的交互;其次,构建与编码器对称的Transformer解码器,对所提取的高阶上下文特征进行多层解码,解码过程中与编码器跳跃连接进行多尺度特征融合,减少下采样造成的空间位置信息丢失;最后,对解码器每个阶段的输出进行监督学习,提升不同尺度特征的抽取质量.基于3个公开的噪声近红外和可见光虹膜数据集NICE.I,CASIA.v4-distance和MICHE-I,与若干包括传统方法、基于卷积神经网络的方法和基于现有Transformer的方法在内的基准方法进行对比实验,实验结果表明,所提方法在E_(1),E_(2),F_(1)和MIOU定量评价指标上均取得了比基准方法更优的分割性能,尤其是在减少噪声的干扰上具有明显的优势.此外,在CASIA.v4-distance数据集上的虹膜识别实验表明,文中方法可以有效地提升虹膜识别的性能,显示了良好的应用潜力.     

一种新的虹膜定位方法

针对现有虹膜定位算法的局限性,提出了一种新的虹膜定位的方法.先对虹膜图像进行预处理,用canny边缘检测算子检测得到虹膜内边缘;再通过二次canny边缘检测对虹膜外边缘进行粗定位,然后分小区域对虹膜外边缘实现多阈值边缘检测,从而获得真实边缘,再利用虹膜内外边缘之间的耦合关系,结合最小二乘法实现虹膜外边缘的定位.实验结果表明,此方法能比较快速准确地定位出虹膜的内外边缘.     

阈值分割算法在医用药瓶溶液杂质检测中的研究

为实现医用药瓶溶液杂质检测,提出一种基于二维Otsu阈值分割法的改进算法,并在Matlab环境中进行仿真。仿真结果表明:相对于传统的二维Otsu法,本文提出的改进方法提高了算法的抗噪性,减少了错误划分,加强了图像的分割效果,且容易实现。     

A new iris segmentation method for non-ideal iris images

Many researchers have studied iris recognition techniques in unconstrained environments, where the probability of acquiring non-ideal iris images is very high due to off-angles, noise, blurring and occlusion by eyelashes, eyelids, glasses, and hair. Although there have been many iris segmentation methods, most focus primarily on the accurate detection with iris images which are captured in a closely controlled environment. This paper proposes a new iris segmentation method that can be used to accurately extract iris regions from non-ideal quality iris images. This research has following three novelties compared to previous works; firstly, the proposed method uses AdaBoost eye detection in order to compensate for the iris detection error caused by the two circular edge detection operations; secondly, it uses a color segmentation technique for detecting obstructions by the ghosting effects of visible light; and thirdly, if there is no extracted corneal specular reflection in the detected pupil and iris regions, the captured iris image is determined as a “closed eye” image.     

Robust and accurate iris segmentation in very noisy iris images

Iris segmentation plays an important role in an accurate iris recognition system. In less constrained environments where iris images are captured at-a-distance and on-the-move, iris segmentation becomes much more difficult due to the effects of significant variation of eye position and size, eyebrows, eyelashes, glasses and contact lenses, and hair, together with illumination changes and varying focus condition. This paper contributes to robust and accurate iris segmentation in very noisy images. Our main contributions are as follows: (1) we propose a limbic boundary localization algorithm that combines K-Means clustering based on the gray-level co-occurrence histogram and an improved Hough transform, and, in possible failures, a complementary method that uses skin information; the best localization between this and the former is selected. (2) An upper eyelid detection approach is presented, which combines a parabolic integro-differential operator and a RANSAC (RANdom SAmple Consensus)-like technique that utilizes edgels detected by a one-dimensional edge detector. (3) A segmentation approach is presented that exploits various techniques and different image information, following the idea of focus of attention, which progressively detects the eye, localizes the limbic and then pupillary boundaries, locates the eyelids and removes the specular highlight.     

Efficient and robust segmentation of noisy iris images for non-cooperative iris recognition

This paper describes the winning algorithm we submitted to the recent NICE.I iris recognition contest. Efficient and robust segmentation of noisy iris images is one of the bottlenecks for non-cooperative iris recognition. To address this problem, a novel iris segmentation algorithm is proposed in this paper. After reflection removal, a clustering based coarse iris localization scheme is first performed to extract a rough position of the iris, as well as to identify non-iris regions such as eyelashes and eyebrows. A novel integrodifferential constellation is then constructed for the localization of pupillary and limbic boundaries, which not only accelerates the traditional integrodifferential operator but also enhances its global convergence. After that, a curvature model and a prediction model are learned to deal with eyelids and eyelashes, respectively. Extensive experiments on the challenging UBIRIS iris image databases demonstrate that encouraging accuracy is achieved by the proposed algorithm which is ranked the best performing algorithm in the recent open contest on iris recognition (the Noisy Iris Challenge Evaluation, NICE.I).     

Real-time iris detection

A real-time algorithm to automatically detect human faces and irises from color images has been developed. A Haar cascade-based algorithm has been applied for simple and fast face detection. The face image is then converted into a gray-scale image. Three types of image processing techniques have been tested to study their effect on the performance of the iris detection algorithm. Then iris candidates are extracted from the valley of the face region. The iris candidates are paired up and the cost of each possible pairing is computed by a combination of mathematical models. The pairing with the lowest cost is considered to be a pair of irises. The algorithm has been tested by quality images from a Logitech camera and noisy images from a Voxx CCD camera. The proposed algorithm has achieved a success rate of 83.60% for iris detection in an open office environment.     

Efficient iris segmentation method in unconstrained environments

Recently, iris recognition systems have gained increased attention especially in non-cooperative environments. One of the crucial steps in the iris recognition system is the iris segmentation because it significantly affects the accuracy of the feature extraction and iris matching steps. Traditional iris segmentation methods provide excellent results when iris images are captured using near infrared cameras under ideal imaging conditions, but the accuracy of these algorithms significantly decreases when the iris images are taken in visible wavelength under non-ideal imaging conditions. In this paper, a new algorithm is proposed to segments iris images captured in visible wavelength under unconstrained environments. The proposed algorithm reduces the error percentage even in the presence of types of noise include iris obstructions and specular reflection. The proposed algorithm starts with determining the expected region of the iris using the K-means clustering algorithm. The Circular Hough Transform (CHT) is then employed in order to estimate the iris radius and center. A new efficient algorithm is developed to detect and isolate the upper eyelids. Finally, the non-iris regions are removed. Results of applying the proposed algorithm on UBIRIS iris image databases demonstrate that it improves the segmentation accuracy and time.     

A knowledge-based approach to the iris segmentation problem

This paper describes a knowledge-based approach to the problem of locating and segmenting the iris in images showing close-up human eyes. This approach is inspired in the expert system’s paradigm but, due the specific processing problems associated with image analysis, uses direct encoding of the “decision rules”, instead of a classic, formalized, knowledge base. The algorithm involves a succession of phases that deal with image pre-processing, pupil location, iris location, combination of pupil and iris, eyelids detection, and filtering of reflections. The development was iterative, based on successive improvements tested over a set of training images. The results that were achieved indicate that this global approach can be useful to solve image analysis problems over which human “experts” have better performance than the present computer-based solutions.     

Reliable algorithm for iris segmentation in eye image

The paper presents a novel algorithm for iris segmentation in eye images taken under visible and near infrared light. The proposed approach consists of the following stages: reflections localization, reflections filling in, iris boundaries localization and eyelids boundaries localization. Here, each of these stages is detailed. Authors’ solution obtained the second rank in the “Noisy Iris Challenge Evaluation – Part I” contest, in which all iris segmentation algorithms submitted to the contest were evaluated and compared.     

Noisy iris segmentation with boundary regularization and reflections removal

The paper presents an innovative algorithm for the segmentation of the iris in noisy images, with boundaries regularization and the removal of the possible existing reflections. In particular, the method aims to extract the iris pattern from the eye image acquired at the visible wavelength, in an uncontrolled environment where reflections and occlusions can also be present, on-the-move and at variable distance. The method achieves the iris segmentation by the following three main steps. The first step locates the centers of the pupil and the iris in the input image. Then two image strips containing the iris boundaries are extracted and linearizated. The last step locates the iris boundary points in the strips and it performs a regularization operation by achieving the exclusion of the outliers and the interpolation of missing points. The obtained curves are then converted into the original image space in order to produce a first segmentation output. Occlusions such as reflections and eyelashes are then identified and removed from the final area of the segmentation. Results indicate that the presented approach is effective and suitable to deal with the iris acquisition in noisy environments. The proposed algorithm ranked seventh in the international Noisy Iris Challenge Evaluation (NICE.I).     

虹膜识别中噪声的检测和处理方法

针对虹膜识别中存在的问题,在识别时应该尽可能将虹膜有效信息提取出来,而将眼皮、睫毛、光点的反射等去除,同时在特征提取中,提出了通过噪卢扩展方法消除噪声信号在编码过程中带来的影响。实验结果表明该文提出的方法有助于提高识别的可靠性和模式分类能力。     

Neural network based biometric personal identification with fast iris segmentation

This paper presents the iris recognition system for biometric personal identification using neural network. Personal identification consists of localization of the iris region and generation of a data set of iris images followed by iris pattern recognition. In this paper, a fast algorithm is proposed for the localization of the inner and outer boundaries of the iris region. Located iris is extracted from an eye image, and, after normalization and enhancement, it is represented by a data set. Using this data set a Neural Network (NN) is used for the classification of iris patterns. The adaptive learning strategy is applied for training of the NN. The results of simulations illustrate the effectiveness of the neural system in personal identification. Recommended by Guest Editor Phill Kyu Rhee. This work was supported by the Near East University. The authors would like to thank Institute of Automation, Chinese Academy of Sciences for providing CASIA iris database. Rahib Hidayat Abiyev was born in Azerbaijan, in 1966. He received the Ph.D. degree in Electrical and Electronic Engineering from Azerbaijan State Oil Academy (old USSR) in 1997. He worked as a Research Assistant at the research laboratory “Industrial intellectual control systems” of Computer-aided control system department. From 1999-present he is working as an Associate Professor at the department of Computer Engineering of Near East University. He is the Chairman of Computer Engineering Department. His research interests are softcomputing, pattern recognition, control systems, signal processing, optimization. Koray Altunkaya was born in Turkey, in 1982. He received the MSc. degree in Computer Engineering from Near East University, North Cyprus in 2007. He is working as an Research Assistant at the research laboratory “Applied Computational Intelligence” of Computer Engineering Department. His research interests are image processing, neural networks, pattern recognition, digital signal processing.