A novel iris localization method and user-specific automatic authentication approach based on feature selection is proposed in this paper. First, two sub-regions, where are nearly not occluded by useless parts such as eyelash eyelid, segmented region of interest (ROI). Second, multi-scale Gabor filters adopted to extract the texture ROI. Third, genetic algorithm (GA) support vector machine (SVM) employed for classification. Through selection, each user has specific index modality. For...

Recent years, people pay more and attention to multi-modal biometric recognition because of its higher security. In this paper, face palmprint are fused in feature level a novel user-specific weighting rule is used adapt different people's need improve accuracy. According respective characteristic palmprint, extraction methods adopted. For face, phase congruency extract energy meanwhile Gabor transformation texture palmprint. Then the two modals weighted according users' specific level....

The δ-generalized labeled multi-Bernoulli (δ-GLMB) filter is an efficient approach for multiobject tracking in case of high clutter density and low detection probability. However, the formulation original δ-GLMB requires that birth filtering known a priori. It inapplicable object appearing from unknown positions. To address this problem, adaptive proposed to detect track objects with position information. This establishes by using measurements at previous three successive times. Simulation...

The problem in parameter selection of least squares support vector machine (LS-SVM) restricts the development LS-SVM, In order to choose optimal parameters LS-SVM automatically, we proposed an improved particle swarm optimization (PSO) algorithm which can not only increase convergent speed but also improve overall searching ability algorithm. PSO increases avoiding local optimum effectively. We use automatically facial expression recognition system. experimental results show that method with...

The interpretation and thickness prediction of seismic thin‐bed is one the difficult problems for geoscientists because band‐limited property data. We propose a new reflectivity inversion method based on matching pursuit algorithm, signal sparse decomposition that widely used in time‐frequency analysis image reconstruction. Different form conventional firstly constructs dictionary by using synthetic waveforms series possible models, then selects optimum models through contrasting all with...

Iris recognition is a kind of novel biometric feature approach which was developed from 1990s and it has attracted more attention because its high accuracy. In this paper, based on researching the existing iris authentication methods, selection practical swarm optimization proposed. We use improved wavelet modulus maximum to locate image extract ROI first. Then we multi-scale Gabor filter feature, can retain features completely reduce computation. At last, GA PSO are used respectively select...

This paper presents a novel lossless and content-based hidden transmission method of biometric images to improve the security secrecy verification. Chaos is used encrypt watermark. The secret keys which called biokey are generated from image as parameter value initial condition chaotic map. digital watermark embedded losslessly in region interest (ROI) palmprint recognize has been destroyed or not regarded user information index for watermarked ROI general public using steganography...

Recent years, the biometric technology attracts more and people's attention. However, single-modal is difficult to meet higher security needs multi-modal gradually becomes mainstream of biometrics. In this paper, face palmprint biological characteristics are used for recognition. Gabor transform Independent Component Analysis (ICA) which rarely in recognition extract their features. Face features can effectively complement each other. The two modals fused feature level with weights...

The problem in accurate and fast face detection tracking video images with complex background is studied this paper. In order to meet both of speed accuracy, a method integrating the skin color segmentation Split up Sparse Network Winnows (SNoW) presented. This kind can not only utilize advantage remove interference background, but also has SNoW accurately classifying feathers. More importantly, inherit their common speed. Based on these advantages, used detect human areas then use way...

The development of graph neural networks (GCN) makes it possible to learn structural features from evolving complex networks. Even though a wide range realistic are directed ones, few existing works investigated the properties and temporal In this paper, we address problem link prediction in propose deep learning model based on GCN self-attention mechanism, namely TSAM. proposed adopts an autoencoder architecture, which utilizes attentional layers capture feature neighborhood nodes, as well...

As a prominent parameter-efficient fine-tuning technique in NLP, prompt tuning is being explored its potential computer vision. Typical methods for visual follow the sequential modeling paradigm stemming from which represents an input image as flattened sequence of token embeddings and then learns set unordered parameterized tokens prefixed to representation prompts task adaptation large vision models. While such has shown great promise, there are two limitations. First, learned cannot model...

In this paper, ridge regression (RR) method is adopted for predicting the formation of microporous aluminophosphate AlPO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> -5 based on a database synthesis, which aims to provide useful guidance rational synthesis materials as well other inorganic crystalline materials. To solve problem class imbalance, four guided resampling methods proposed in our previous work are employed improving...

Performance analysis is an important guarantee for improving teaching and learning methods. The results of performance are the data support teachers to adjust plans design content guide students their plans. method student achievement based on decision tree algorithm proposed. This realizes effective students' grades through three stages: collection preprocessing, construction grade model, prediction. verified by multiple majors in a university, experimental show that this has good effect.