CASIA Database Research Articles

Iris recognition using combined support vector machine and Hamming distance approach

Iris based authentication system is essentially a pattern recognition technique that makes use of iris patterns, which are statistically unique, for the purpose of personal identification. In this study, a novel method for recognition of iris patterns is considered by using a combination of support vector machine and Hamming distance. The zigzag collarette area of the iris is selected for iris feature extraction because it captures the most important areas of iris complex pattern and higher recognition rate is achieved. The proposed approach also used parabola detection and trimmed median filter for the purpose of eyelid and eyelash detection & removal, respectively. The proposed method is computationally effective as well as reliable with a recognition rate of 99.91% and 99.88% on CASIA and Chek image database respectively.

Gait-Based Person Identification Robust to Changes in Appearance

The identification of a person from gait images is generally sensitive to appearance changes, such as variations of clothes and belongings. One possibility to deal with this problem is to collect possible subjects' appearance changes in a database. However, it is almost impossible to predict all appearance changes in advance. In this paper, we propose a novel method, which allows robustly identifying people in spite of changes in appearance, without using a database of predicted appearance changes. In the proposed method, firstly, the human body image is divided into multiple areas, and features for each area are extracted. Next, a matching weight for each area is estimated based on the similarity between the extracted features and those in the database for standard clothes. Finally, the subject is identified by weighted integration of similarities in all areas. Experiments using the gait database CASIA show the best correct classification rate compared with conventional methods experiments.

Hashing Based Fast Palmprint Identification for Large-Scale Databases

In this paper, we present two hashing based techniques for fast palmprint identification. We first propose three properties required by a hash function and then introduce our first fast identification method based on orientation pattern (OP) hashing. We give the definition of OP and demonstrate that it meets all the requirements of the hash function, and thus, would be appropriate for hashing based fast palmprint identification. We then introduce the second fast identification method based on principal orientation pattern (POP) hashing. Because the POPs are constructed using more stable orientation features, POP hashing can find the target template more quickly thus causing earlier termination of the identification process. We evaluate our methods on the Hong Kong PolyU large-scale database (9667 palms) and the CASIA palmprint database (600 palms) plus a synthetic database (100 000 palms). Experimental results show that, on the Hong Kong PolyU large-scale database, the speedups of OP hashing and POP hashing over brute-force search are 16.93 and 19.91, respectively, and the identification accuracy is slightly higher. While on the CASIA database plus the synthetic database, the speedups of OP hashing and POP hashing are 8.03 and 15.67, respectively, and the identification accuracy almost remains the same. Results also show that, in terms of accuracy, our methods are comparable to several state-of-the-art palmprint identification approaches, while in terms of speed, our methods are much faster.

A NEW APPROACH FOR FACE-IRIS MULTIMODAL BIOMETRIC RECOGNITION USING SCORE FUSION

In this paper, a new approach based on score level fusion is presented to obtain a robust recognition system by concatenating face and iris scores of several standard classifiers. The proposed method concatenates face and iris match scores instead of concatenating features as in feature-level fusion. The features from face and iris are extracted using local and global feature extraction methods such as PCA, subspace LDA, spPCA, mPCA and LBP. Transformation-based score fusion and classifier-based score fusion are then involved in the process to obtain, concatenate and classify the matching scores. Different fusion techniques at matching score level, feature level and decision level are compared with the proposed method to emphasize improvement and effectiveness of the proposed method. In order to validate the proposed scheme, a combined database is formed using ORL and BANCA face databases together with CASIA and UBIRIS iris databases. The results based on recognition performance and ROC analysis demonstrate that the proposed score level fusion achieves a significant improvement over unimodal methods and other multimodal face-iris fusion methods.

Iris Texture Analysis for Security Systems

Biometric security is highly reliable and secure system. Proposed work, efficient methods for pupil detection and wavelet transformation with five-level decomposition for feature extraction are proposed and which results with accurate feature vector are stored as bits and then processes for identification and verification. Statistical performance evaluation using parameters and classifier used hamming distance for matching the patterns efficiently with stored database .We use CASIA database.

Assessment of geometric features for individual identification and verification in biometric hand systems

This paper studies the reliability of geometric features for the identification of users based on hand biometrics. Our methodology is based on genetic algorithms and mutual information. The aim is to provide a system for user identification rather than a classification. Additionally, a robust hand segmentation method to extract the hand silhouette and a set of geometric features in hard and complex environments is described. This paper focuses on studying how important and discriminating the hand geometric features are, and if they are suitable in developing a robust and reliable biometric identification. Several public databases have been used to test our method. As a result, the number of required features have been drastically reduced from datasets with more than 400 features. In fact, good classification rates with about 50 features on average are achieved, with a 100% accuracy using the GA-LDA strategy for the GPDS database and 97% for the CASIA and IITD databases, approximately. For these last contact-less databases, reasonable EER rates are also obtained.

Fast self-generation voting for handwritten Chinese character recognition

In this paper, a fast self-generation voting method is proposed for further improving the performance in handwritten Chinese character recognition. In this method, firstly, a set of samples are generated by the proposed fast self-generation method, and then these samples are classified by the baseline classifier, and the final recognition result is determined by voting from these classification results. Two methods that are normalization-cooperated feature extraction strategy and an approximated line density are used for speeding up the self-generation method. We evaluate the proposed method on the CASIA and CASIA-HWDB1.1 databases. High recognition rate of 98.84 % on the CASIA database and 91.17 % on the CASIA-HWDB1.1 database are obtained. These results demonstrate that the proposed method outperforms the state-of-the-art methods and is useful for practical applications.

Efficient Iris Recognition Algorithm Using Method of Moments

This paper presents an efficient biometric algorithm for iris recognition using Fast Fourier Transform and moments. Biometric system provides automatic identification of an individual based on a unique feature or characteristic possessed by the individual. The Fast Fourier Transform converts image from spatial domain to frequency domain and also filters noise in the image giving more precise information. Moments are area descriptors used to characterize the shape and size of the image. The moments values are invariant to scale and orientation of the object under study, also insensitive to rotation and scale transformation. At last Euclidean distance formula is used for image matching. The CASIA database clearly demonstrates an efficient method for Biometrics. As per experimental result,the algorithm is achieving higher Correct Recognition Rate.

Iris recognition based on bidimensional empirical mode decomposition and fractal dimension

As the demand for information security increases, more attention is being paid to biometrics-based, automated personal identification. One of the most promising current biometric techniques is based on the human iris. This paper attempts to detect shape information from the iris by analyzing local intensity variations of an iris image. The methodology involves extraction of iris features using bidimensional empirical mode decomposition (BEMD) and fractal dimension. After the preprocessing procedure, the normalized effective iris image is decomposed into 2D intrinsic mode function (IMF) components at different spatial frequencies by bidimensional empirical mode decomposition. Then the texture features of each intrinsic mode function image are obtained via the differential box-counting method. To evaluate the efficacy of the proposed approach, three different similarity measures used in recognition are adopted. The experimental results using the CASIA and ICE iris databases show that the schema presented achieves promising results for iris recognition.

Gait Recognition using Shoulder Body Joint

This paper introduces a new gait recognition approach based on shoulder body joint using in intersection point. This approach is tested on a CASIA database of video sequences, corresponding 10 different subject of people. Obtained better recognition rate as compared to other methods. Consider two component of human body. The first component is shoulder and second component is feet. The second component is subdivided into two parts i.e. toe and heel of both right and left leg. Our proposed method increase the matching accuracy with 98 and 100 percent.

A probabilistic image-weighting scheme for robust silhouette-based gait recognition

Many gait recognition methods use silhouettes as a feature due to their simplicity and effectiveness. However, silhouette-based gait recognition algorithms have the drawback of performance degradation when the silhouette images are corrupted. To solve this problem, this paper proposes a new gait representation method by emphasizing the noise-free silhouettes while suppressing the corrupted ones. The probabilistic support vector machine (PSVM) is employed to weigh the silhouette images according to quality and to construct a new gait representation for robust recognition. Experiments are conducted with the CASIA and SOTON databases, and the proposed method makes silhouette-based gait recognition as reliable biometrics.

A Supervised Hybrid Methodology for Pose and Illumination Invariant 3D Face Recognition

D face recognition systems encounter difficulties in recognizing faces with illumination variations. The depth map of the 3D face data has the potential to handle the variation in illumination of face images. The view variations are handled by using the moment invariants. Moment Invariants are used as rotation invariant features of the face image. For feature matching an efficient fuzzy-neural technique is proposed. The PCA components of normalized depth map and Moment invariants on mesh images are used successfully to implement a fuzzy neural network based fully automatic 3D face recognition system. The system is evaluated on the 3D face databases, the CASIA database. The proposed system provides recognition accuracy that is resulted in to an efficient 3D face recognition system. Keywords-Neural Network (FNN), 3D-Face Recognition, Depth Map, Moment Invariants.

The Comparison of Iris Recognition using Principal Component Analysis, Log Gabor and Gabor Wavelets

an ever growing emphasis on security systems, automated personal identification based on biometrics has been getting extensive focus in both research and practical over the last decade. The methods for iris recognition mainly focus on feature representation and matching. As we known traditional iris recognition method is using Gabor Wavelet features, the iris recognition is performed by a 256 byte iris code, which is computed by applying Gabor wavelets to a given portion of iris. Log Gabor wavelet based features are invariant to changes in brightness and illumination whereas techniques like principal component analysis can produce spatially global features. The goal of this paper is to compare feature extraction algorithm based on PCA, Log Gabor Wavelet and Gabor Wavelet. We use these methods to generate feature vectors that could represent iris efficiently. Conclusions based on comparisons can provide useful information for further research. Performance of these algorithms is analyzed using CASIA database.

Slow Feature Analysis for Human Action Recognition

Slow Feature Analysis (SFA) extracts slowly varying features from a quickly varying input signal. It has been successfully applied to modeling the visual receptive fields of the cortical neurons. Sufficient experimental results in neuroscience suggest that the temporal slowness principle is a general learning principle in visual perception. In this paper, we introduce the SFA framework to the problem of human action recognition by incorporating the discriminative information with SFA learning and considering the spatial relationship of body parts. In particular, we consider four kinds of SFA learning strategies, including the original unsupervised SFA (U-SFA), the supervised SFA (S-SFA), the discriminative SFA (D-SFA), and the spatial discriminative SFA (SD-SFA), to extract slow feature functions from a large amount of training cuboids which are obtained by random sampling in motion boundaries. Afterward, to represent action sequences, the squared first order temporal derivatives are accumulated over all transformed cuboids into one feature vector, which is termed the Accumulated Squared Derivative (ASD) feature. The ASD feature encodes the statistical distribution of slow features in an action sequence. Finally, a linear support vector machine (SVM) is trained to classify actions represented by ASD features. We conduct extensive experiments, including two sets of control experiments, two sets of large scale experiments on the KTH and Weizmann databases, and two sets of experiments on the CASIA and UT-interaction databases, to demonstrate the effectiveness of SFA for human action recognition. Experimental results suggest that the SFA-based approach (1) is able to extract useful motion patterns and improves the recognition performance, (2) requires less intermediate processing steps but achieves comparable or even better performance, and (3) has good potential to recognize complex multiperson activities.

3D Face Recognition Method Based on Cascade Classifier

A method for 3D face recognition based on Cascade Classifier algorithm is proposed in this paper. First, 3D point data is mapped to X-Y planar by using iteration interpolation method, and the range image is obtained. Then the principal component analysis (PCA) is used to find a low dimensional feature face space and then we achieve match score in accordance with the nearest neighbor rule. Finally, the remaining part faces are verified based on Hausdorff distance algorithm. The simulation experiment is done on CASIA database, the result shows that the performance of our algorithm has some improvement compared with other algorithms.

3D Facial Gender Classification Based on Multi-angle LBP Feature

Facial gender classification is a challenging topic,and it s still not perfect until now.In this paper,we propose a series of methods of gender classification based on three-dimension faces.Automatic front-pose adjustment is needed through local region iterative closest point(ICP) registration firstly;then we do pitching rotating and extract multi-angle LBP features from depth thumbnail map in di?erent viewing angles;at last,we use support vector machine(SVM) classifier to do training and prediction.This algorithm has been experimented on CASIA database,and for the neutral faces in this database,we can get a highest correct classification rate of 98.374%.

Multi-Modal Biometric Template Security: Fingerprint and Palmprint Based Fuzzy Vault

Multi-biometric system stores multiple templates for the same user corresponding to the different biometric sources. Infallible security should be provided to the stored biometric templates as biometric is not revocable. In this work, multi-modal biometric template security for palmprint and fingerprint is proposed which is based on the fuzzy vault generation. At first, the preprocessing steps are applied and subsequently, the features are extracted and combined. For recognition, we match the feature vectors of images. The multi-modal biometric template along with the input key are used to generate the fuzzy vault. In the decoding process, the template is given as input and is combined with the stored fuzzy vault to generate the corresponding final key. The experimentation is carried out using CASIA database for palmprint and FVC 2004 database for fingerprint. The evaluation metrics have FMR and FNMR value parameters.

A FRAMEWORK TOWARDS USING EIGEN IRIS, MINUTIAE THUMB AND DNA SEQUENCE FEATURES FOR PERSONAL IDENTIFICATION

Biometric systems based on a single physiological or behavioral characteristic may not be able to identify a person correctly. This paper presents an efficient and reliable multimodal biometric identification system which is based on minutiae thumb, Eigen iris and DNA sequence features. In this method compressed form of Short tandem repeat (STR) part of DNA sequence, compressed thumbprint and compressed iris image (Eigen values) of a person are used for further identification of an individual. Therefore, personal identification including identical twins and dead person's cases will become easier in using this method. Our technique will correctly identify a person (living or dead) on the basis of his thumbprint, iris image and DNA sequence features. We have used thumbprint and iris images for identifying a live person. But, for identifying a dead person we have used STR part of DNA sequence. We tested our method for 100 samples of thumbprints and iris images of CASIA database and we found that our multimodal method is able to correctly identify each and every individual including identical twin on the basis of thumbprints and iris images. For identifying a dead person the compressed form of DNA sequence of that person is used.

AN EFFICIENT EIGEN VALUES BASED TECHNIQUE FOR ONLINE IRIS IMAGE COMPRESSION AND IDENTIFICATION

This paper introduces an efficient Eigen values based technique for online iris image compression and identification of a human including the case of identical twins. The iris image is extracted after removing the pupil, eye brow, skin and other noise disturbances from an actual image. The extracted iris image is divided into different blocks of size of 16 × 16. Now, Eigen values are calculated for each block and these Eigen values are stored in the smart card memory for further identification. Therefore, when checking if two iris images are identical or not, all we need is to compare the stored Eigen values with online calculated Eigen values. If two iris images have the same Eigen values, this means that both iris images belong to the same person. In our research, we have concluded that iris images of different persons have different Eigen values, including the case of identical twins. We conducted experiments on CASIA and Multimedia University iris image databases and we found that our Eigen Values Based Iris Image Identification Technique is giving 99.99% accuracy for the same image of identical twins and individuals. The implementation leads us to believe that our method is giving the best matching result in the case of identical twins and individuals. It is an efficient, secure and economically feasible approach for online personal identification.

An Algorithm for Human Iris Template Matching

Iris recognition systems are widely used in biometrical based technology for personal identification and verification. Many systems have been implemented and their performances have been tested; each of them has its own advantages and disadvantages. In order to get over the drawbacks of these systems mixed techniques were described to implement an iris recognition system. The circular iris and pupil were segmented using threshold concepts and a canny edge detector followed by Circular Hough transform (CHT). Localized iris was normalized by Daugman's rubber sheet model. The coding method based on 1D log-Gabor transform is used to extract the discriminating features. Finally Hamming Distance (HD) operator was used in the template matching process. The proposed system was tested on CASIA database version-1. The results obtained for the hamming distance shows a significant improvement in HD operator as well as the processing time.