Review of pixel-level image fusion

Image fusion can be performed at different levels:signal,pixel,feature and symbol levels.Almost all image fusion algorithms developed to date fall into pixel level.This paper provides an overview of the most widely used pixel-level image fusion algorithms and some comments about their relative strengths and weaknesses.Particular emphasis is placed on multiscale-based methods.Some performance measures practicable for pixel-level image fusion are also discussed.At last,prospects of pixel-level image fusion ar...     

Multi-Image Restoration Method Combined with Total Generalized Variation and l_p-Norm Regularizations

Image restoration is an important part of various applications, such as computer vision, robotics and remote sensing. However, recovering the underlying structures of the latent image contained in multi-image is a challenging problem because of the need to develop robust and fast algorithms. In this paper, a novel problem formulation for multi-image restoration problem is proposed. This novel formulation is composed of multi-data fidelity terms and a composite regularizer. The proposed regularizer consists of total generalized variation(TGV)and lp-norm. This multi-regularization method can simultaneously exploit the consistence of image pixels and promote the sparsity of natural signals. To deal with the resulting problem, we derive and implement the solution using alternating direction method of multipliers(ADMM). The effectiveness of our method is illustrated through extensive experiments on multi-image denoising and inpainting. Numerical results show that the proposed method is more efficient than competing algorithms, achieving better restoration performance.     

Incremental Supervised Subspace Learning for Face Recognition

Subspace learning algorithms have been well studied in face recognition. Among them, linear discriminant analysis (LDA) is one of the most widely used supervised subspace learning method. Due to the difficulty of designing an incremental solution of the eigen decomposition on the product of matrices, there is little work for computing LDA incrementally. To avoid this limitation, an incremental supervised subspace learning (ISSL) algorithm was proposed, which incrementally learns an adaptive subspace by optimizing the maximum margin criterion (MMC). With the dynamically added face images, ISSL can effectively constrain the computational cost. Feasibility of the new algorithm has been successfully tested on different face data sets.     

Synthetically Evaluation System for Multi-source Image Fusion and Experimental Analysis

Introduction The image fusion technology produces a singleimage from a set of input images. The fused imageshould have more complete information which ismore useful for human or machine perception.With the merits of high degree of system explo-ration, automation, availability, reliability, capa-bility and low cost, it boasts broad application per-spective in many fields such as computer vision,automatic object detection, image processing, par-allel and distributed processing, robotics and re-m…     

Spectrum Feature Retrieval and Comparison of Remote Sensing Images Using Improved ISODATA Algorithm

Due to the large quantities of data and high relativity of the spectra of remote sensing images, K-L transformation is used to eliminate the relativity. An improved ISODATA (Interative Self-Organizing Data Analysis Technique A) algorithm is used to extract the spectrum features of the images. The computation is greatly reduced and the dynamic arguments are realized. The comparison of features between two images is carried out, and good results are achieved in simulation.     

Efficient Recognition of Human Faces from Video in Particle Filter

Introduction As an active research in computer vision andimage understanding, face recognition from videohas got wide applications, such as human-comput-er interface, video surveillance, ATM and videocommunications[1]. So far, there are many litera-tures on face recognition from video. Many archi-tectures about dynamic face recognition were pro-posed in those literatures. Tracking and recogni-tion were performed separately in Ref.[2]. Thelimitation in this method is that tracking andrecognit…     

Mean shift registration algorithm for dissimilar sensors

The mean shift registration (MSR) algorithm is proposed to accurately estimate the biases for multiple dissimilar sensors. The new algorithm is a batch optimization procedure. The maximum likelihood estimator is used to estimate the target states, and then the mean shift algorithm is implemented to estimate the sensor biases. Monte Carlo simulations show that the MSR algorithm has significant improvement in performance with reducing the standard deviation and mean of sensor biased estimation error compared with the maximum likelihood registration algorithm. The quantitative analysis and the qualitative analysis show that the MSR algorithm has less computation than the maximum likelihood registration method.