This paper proposes a dual-biometric-modality personal identification system, which has both outstanding identification performance and effective anti-spoofing property. This dual-modality system represents the first effort to fuse the conventional fingerprint with a novel biometric modality--Electroencephalogram (EEG). The fusion of these two modalities is achieved at the matching score level. Experimental results show that the highest identification performance is obtained in the dual-modality system, when its performance is compared with those of the systems using these two modalities separately. Furthermore, it is demonstrated that EEG is an appealing complementary modality to enhance the anti-spoofing capability of conventional biometrics-based identification systems.

No abstract

The problem of noninvasive computing the epicardial surface potentials from torso surface potentials constitutes one form of the inverse problem of ECG, which can be acted as a regression problem with multi-input and multi-output. In this study, the SVR method is invoked to predict the inverse solutions, which compared with the common regularization methods. To build an effective SVR model, the hyper-parameters of SVR are set carefully by using the grid search optimization method. The experiment results shows that SVR method is an effective way for solving the inverse ECG problem, which can reconstruct more accurate epicardial surface potentials distribution than the common regularization method, such as Tikhonv method and LSQR method.

No abstract

Accurately object searching plays an important role in computer vision. Retrieving and locating target objects in images are object searching's two sub-tasks. Aiming to promote the precision and recall of object searching, selecting appropriate image representation methods is the core issues. The representation method needs to provide enough discriminative features. Our approach adopts locality sensitive hashing method to extract enough sift features. The extracted features contain inliers and outliers. In order to distinguish them, random context confidence scores of features are computed. Our algorithm offers 3 benefits:1) A novel partition method is adopted to divide images. It is easy to be parallelized during computing contexts.2) A novel random points selecting method is adopted to avoids ill-defined boundary for target objects; 3) Multiple target objects in one image can be located by clustering all the features of each image with their coordinates. The experiment on a challenging Belgalogo dataset highlights the performance of our approach.