Yuqing He

Optimization of phase mask-based iris imaging system through the optical characteristics

Iris recognition is the most reliable method in personal identification. However, the current fixed-focus iris imaging system has small depth of field (DOF), which limits the wide application of the iris recognition system. This paper presents the design method and optimization of a phase mask based iris imaging system. Through wavefront coding, it can extend the DOF and enhance the convenience of iris image acquisition. Through analyzing the modulation transfer function and optical parameters of the cubic phase mask, we can get the wavefront coding iris imaging system’s optimal parameter and it’s structure. Experimental results show that the cubic phase mask based iris imaging system has larger DOF and better imaging performance.

A review of advances in iris image acquisition system

Iris recognition is a high-precision biometric identification technology with the advantages of uniqueness, stability, non-invasive. Iris images quality affect the performance of the recognition algorithms. The ease of use and robustness of the recognition system is also affected by the image acquisition method, so iris image acquisition plays an important role in the whole system. Based on the basic principles of iris image acquisition, this paper gives the current advances of the iris recognition system. Describes and analyzes the typical commercial products of iris image acquisition system, including the operating range, capture volume, illumination mode, etc.. According to the bottleneck of the current iris image acquisition and recognition system, major research issues in the area of iris image acquisition are presented and analyzed, such as the stand-off system, variety of illumination mode, etc. At last, gives the development trend and future work of the iris image acquisition system.

Super resolution reconstruction and recognition for iris image sequence

As a non-invasive and stable biometric identification method, iris recognition is widely used in safety certification. In large scenes or long-distance conditions, the iris images acquired may has low resolution. Lack of information in these images or videos affects the performance of the iris recognition greatly. In this paper, we proposed a scheme of super resolution to reconstruct high-resolution images from low-resolution iris image sequences. The proposed scheme applies an improved iterated back projection algorithm to reconstruct high-resolution images and does not have a restriction on the numbers of base images. We simulated our method and conducted experiments on a public database. The results show that the reconstructed high-resolution iris image provides enough pixels which contain sufficient texture information for recognition. Lower Equal Error Rate is achieved after the robust super resolution iris image reconstruction.

Combining Multiple Color Components for Efficient Visible Spectral Iris Localization

Iris localization is the prerequisite for the precise iris recognition. Compared with near-infrared iris images, the visible spectral iris images may have more fuzzy boundaries, which impair the iris detection. We can use multiple color components of different color spaces to realize the visible spectral iris localization. Firstly, the sclera is segmented and eyelids are located on \( \alpha \) component image through contrast adjustment and polynomial fitting. Secondly, morphological processing and CHT (Circular Hough Transform) is applied to localize the limbic boundary on R component image. Similarly, the pupillary boundary is localized on R component image and \( \alpha \) component image. Experimental results on visible spectral iris image dataset indicate that the proposed method has good performance on iris localization.

Design of an Embedded Multi-biometric Recognition Platform Based on DSP and ARM

Thedual-core embedded system can make the system have higher efficiency of simultaneous running in the recognition algorithm and controlling the peripheral equipments. This paper presents a study of an embedded multi-biometric recognition system based on DSP and ARM which can realize the face and iris image acquisition, recognition, datastorage and input/output control. The ARM is used as a host to communicate with peripherals. The DSP performs the multi-biometric image acquisition and recognition. The host port interface (HPI) is used to implement the communication between DSP and ARM. We design the HPI strobe signal and the hardware device driver based on the embedded Linux to realize the data exchange and communication. Experimental results show that the dual-core embedded system has greater storage capacity and higher interactive ability.

An Embedded Self-adaptive Iris Image Acquisition System in a Large Working Volume

Iris image acquisition is a key step in the iris recognition. Usually most of current systems have a short working volume and users need to cooperate in the specific range, which limits the system application. In this paper, we designed an embedded self-adaptive iris image acquisition system using a single camera with a large working volume. It can capture the user’s iris in the distance from 0.3 meter to 1.1 meter. A variable zoom camera is co-located in a pan-tilt-unit (PTU) for face detection and iris image acquisition. Combining the face detection and eye location, the system can center and zoom the camera for eyes. The micro controller unit (MCU) controls the peripheral components including PTU, camera, distance sensor, etc. The DSP is used to realize the algorithm and communicate with MCU. Experimental results show the proposed system can capture high-quality iris images efficiently.

Plantar Pressure Data Based Gait Recognition by Using Long Short-Term Memory Network

As a kind of continuous time series, plantar pressure data contains rich contact of time information which has not been fully utilized in existing gait recognition methods. In this paper, we proposed a new gait recognition method based on plantar pressure data with a Long Short-Term Memory (LSTM) network. By normalization and dimensionality reduction, the raw pressure data was converted to feature tensor. Then we feed the LSTM network with the feature tensors and implement classification recognition. We collected data from 93 subjects of different age groups, and each subjects was collected 10 sets of pressure data. The experiment results turn out that our LSTM network can get high classification accuracy and performs better than CNN model and many traditional methods.

Visible Spectral Iris Segmentation via Deep Convolutional Network

Iris segmentation is the prerequisite for the precise iris recognition. Visible spectral iris images may result in lower segmentation accuracy due to noise interference. We use deep learning method to segment the iris region in visible spectral iris images. A deep convolution neural network is designed to extract the eye features and segment the iris, pupil, sclera and background. It’s an end-to-end model which requires no further processing. We collect the eye images and manually mask different part of the eye to establish the visible spectral iris dataset for training and testing. The proposed method was trained based on DeepLab framework. Experimental results show that the proposed method has efficiency on iris segmentation.

Texture Enhancement of Iris Images Acquired under Natural Light

The iris image acquired under natural light may be degraded by non-uniform illumination, which results in the iris texture’s low resolution and low contrast. The recognition accuracy may be affected. This paper describes a method for enhancing the iris textures on the V channel of HSV space. The enhancement has two steps. First, the image is divided into small blocks and luminance enhancement is carried out in each block by using nonlinear transfer function and bilinear interpolation. Secondly, contrast enhancement by multi-scale Gaussian convolution is applied to improve the quality of the image. We test the proposed method on UBIRIS.v2 database. Experimental results show that the proposed method has better texture enhancement performance and can achieve higher recognition accuracy.

Automatic target locating system through cooperative dual-field imaging

This paper proposes an automatic targeting locating system based on dual-field imaging to improve the stability of light weapons. The system consists of a wide field of view (WFOV) camera and a narrow field of view (NFOV) camera. The WFOV camera searches the pedestrian in the scenery, the other camera tracks the pedestrian and aims it accurately. Video signal is send to the processing unit PC and control signal is send back to the imaging system. This automatic target tracking algorithm is integrated by Adaboost and Median-Flow algorithm. It is used to track the pedestrians and locate the head of the target. Experiment results show that the dual-field imaging system and proposed algorithm has robust target tracking performance.