International Journal of Pattern Recognition and Artificial Intelligence

This paper proposed an algorithm for a deep learning network for identifying circular traffic lights (CTL-DNNet). The sample labeling process uses translation to increase the number of positive samples, and the similarity is calculated to reduce the number of negative samples, thereby reducing overfitting. We use a dataset of approximately 370000 samples, with approximately 20000 positive samples and approximately 350000 negative samples. The datasets are generated from images taken at the Beijing Garden Expo. To obtain a very robust method for the detection of traffic lights, we use different layers, different cost functions and different activation functions of the depth neural network for training and comparison. Our algorithm has evaluated autonomous vehicles in varying illumination and gets the result with high accuracy and robustness. The experimental results show that CTL-DNNet is effective at recognizing road traffic lights in the Beijing Garden Expo area.

Text-independent handwriting identification methods require that features such as texture are extracted from lengthy document image; while text-dependent handwriting identification methods require that the contents of the documents being compared are identical. In order to overcome these confinements, this paper presents a novel Chinese handwriting identification technique. First, Chinese characters are segmented from handwriting document, then keywords are extracted based on matching and voting of local features of character. Then the same-content keywords are used to build training sets, and these training sets of two documents are compared. Because the keywords are similar to signature, the handwriting identification problem is transformed into signature verification problem. Experiments on HIT-MW, HIT-SW and CASIA show this method outperforms many text-independent handwriting identification methods.

This paper presents a novel approach to quadrangulate a sequence of meshes with global optimization method, generating consistent quad meshes expressing global topological and geometric characteristics during animation. The key contribution is that our method first proposes a grouping strategy to extract the geometric and topological features of the sequence in animation quadrangulation. Animation sequence is firstly divided into groups through picked key frames. Then analyze the local deformation of each group and detect the surface feature for each reference frame, inducing a set of hard and soft constraints for each key frame of every subgroup, next smooth the cross field of each reference frame and combine them through singularities for the whole sequence, finally compute a field guided global parametrization and quad meshes extraction with an off-the-shelf method. It is effective and easy to implement, and experiment results and comparisons show that our method achieves better quality.

A novel image matching algorithm based on both Taguchi method and spatial clustering is proposed to optimize the Scale Invariant Feature Transform (SIFT) matching results. To improve the matching accuracy, adaptive spatial clustering is used. What is more, in order to get the fitting parameters to balance matching accuracy and quantity, Taguchi method is adopted to optimize the key parameter combination including the ratio threshold of Euclidean distance and the constrain parameters in the process of adaptive spatial clustering. Moreover, signal-to-noise ratio (SNR) results are analyzed by variance to get the effect factor which is taken as the basis for the selection of optimized parameters. The optimum parameters combination is obtained eventually. The final experimental results show that the matching quality based on SIFT feature are improved significantly.

This paper proposes an efficient and fast iris localization method. It uses support vector machine learning of iris features that represent closed outer and inner iris boundaries encompassing a low-intensity region. In addition, depending on the location of the iris in an eye image, an iris detection method is proposed based on three sub-datasets of eye images (middle, right, and left sub-datasets) with different iris features. The proposed method is implemented using fast sliding window and fast computation of the iris detection score with binary features. Compared with state-of-the-art methods, experimental results show that the proposed method is twice as fast and has comparable accuracy, even when factoring in head rotation, glasses, and highlights.

Automatic retinal image registration is still a great challenge in computer aided diagnosis and screening system. In this paper, a new retinal image registration method is proposed based on the combination of blood vessel segmentation and scale invariant feature transform (SIFT) feature. The algorithm includes two stages: retinal image segmentation and registration. In the segmentation stage, the blood vessel is segmented by using the guided filter to enhance the vessel structure and the bottom-hat transformation to extract blood vessel. In the registration stage, the SIFT algorithm is adopted to detect the feature of vessel segmentation image, complemented by using a random sample consensus (RANSAC) algorithm to eliminate incorrect matches. We evaluate our method from both segmentation and registration aspects. For segmentation evaluation, we test our method on DRIVE database, which provides manually labeled images from two specialists. The experimental results show that our method achieves 0.9562 in accuracy (Acc), which presents competitive performance compare to other existing segmentation methods. For registration evaluation, we test our method on STARE database, and the experimental results demonstrate the superior performance of the proposed method, which makes the algorithm a suitable tool for automated retinal image analysis.

In this paper, a parametric geometry model based on elliptic and conic surfaces was developed for bioprosthetic heart valve (BHV) simulation. The valve material was modeled by a hyperelastic nonlinear anisotropic solid model. Different suture densities could be substituted by various bonded points between artery vessel and the leaflets as boundary conditions in the computational modeling. Besides these two assumptions that dynamic structure (DS) and fluid–structure interaction (FSI) both shared, the latter need incompressible viscous Newton fluid model to depict bloodstream passing through the BHV. Immersed boundary (IB) method was introduced to solve the FSI simulation. In addition, the DS analysis applied transvalvular pressure on the valve while FSI had left ventricular pressure on fluid inlet as initials. There was inconsistency between the moments of the maximum deformation and the maximum loading in both simulations. Although a similar trend of deformation lagging the load was viewed, the extent of delay in FSI was much smaller compared with that in DS simulation. The deformed profiles in cross-sectional views were shown in one picture to illustrate different dynamic responses caused by distinct assumptions. Percent of open area at the moments when the maximum deformation occurred was defined to show which calculation achieved better approximation for precise hemodynamics. Fixed point was given as boundaries between BHV and artery in the modeling part. Calculations showed that the more the fixed points in this bonded contact, the lower the principal stress was. The maximum shear stress showed a different trend. It had a different trend. Stress concentration in the conjunction area made it high-risk to be teared. Different suture densities had significant impaction in FSI simulations. With that analysis our work achieved a more comprehensive simulation to describe true hemodynamics of a BHV implanted in artery. The artery vessel had particular dynamic response under such assumptions, gradient existed in the maximum principal stress distribution diagram, from inner wall through which blood passing to the outer wall. Results showed a large suture density was suggested in BHV implantation.

Objective: To investigate the effectiveness of combined laser needle-knife and massage on cervical spondylotic vertebral arteriopathy patients. Summary of background data: With the recent rise of electronic businesses, the incidence of cervical spondylosis has also risen rapidly. Methods: Cervical spondylotic vertebral arteriopathy patients were treated using laser needle-knife with massage, and compared to patients who only received the massage. A 3D anatomical and hemodynamic model was developed. Results: The symptomatic and functional overall scores were reduced by 71.43% after the combined treatment. Results from the finite element analysis indicated that the maximum flow rate of the left vertebral artery was improved by 47.52% and the right was improved by 38.67%. Conclusion: A combined treatment of cervical spondylotic vertebral arteriopathy using laser needle-knife and massage is an effective approach with a therapeutic mechanism closely related to hemodynamics.

Nowadays, department stores and online merchants usually develop some price promotion strategies to attract customers and increase their purchase intention. Therefore, it is significant for customers to pick out attractive products and obtain the maximum discount rate. Admittedly, the skyline query is a most useful tool to find out attractive products. However, it does little to help select the product combinations with the maximum discount rate. Motivated by this, we identify an interesting problem, a most preferential skyline product (MPSP) combination discovering problem, which is NP-hard, for the first time in the literature. This problem aims to report all skyline product combinations having the maximum discount rate. Since the exact algorithm for the MPSP is not scalable to large or high-dimensional datasets, we design an approximate algorithm that guarantees the accuracy of the results. The experiment results demonstrate the efficiency and effectiveness of our proposed algorithms.

Differential Evolution (DE) is a population-based Evolutionary Algorithm (EA) for solving optimization problems over continuous spaces. Many optimization problems are constrained and have a bounded search space from which some vectors leave when the mutation operator of DE is applied. Therefore, it is necessary the use of a boundary constraint-handling method (BCHM) in order to repair the invalid mutant vectors. This paper presents a generalized and improved version of the Centroid BCHM in order to keep the search within the valid ranges of decision variables in constrained numerical optimization problems (CNOPs), which has been tested on a robust and comprehensive set of experiments that include a variant of DE specialized in dealing with CNOPs. This new version, named Centroid $K+1$, relocates the mutant vector in the centroid formed by K random vectors and one vector taken from the population that is within or near the feasible region. The results show that this new version has a major impact on the algorithm’s performance, and it is able to promote better final results through the improvement of both, the approach to the feasible region and the ability to generate better solutions.

Topological link prediction is the task of assessing the likelihood of new future links based on topological properties of entities in a network at a given time. In this paper, we introduce a multistrain bacterial diffusion model for link prediction, where the ranking of candidate links is based on the mutual transfer of bacteria strains via physical social contact. The model incorporates parameters like efficiency of the receiver surface, reproduction rate and number of social contacts. The basic idea is that entities continuously infect their neighborhood with their own bacteria strains, and such infections are iteratively propagated on the social network over time. The probability of transmission can be evaluated in terms of strains, reproduction, previous transfer, surface transfer efficiency, number of direct social contacts i.e. neighbors, multiple paths between entities. The value of the mutual strains of infection between a pair of entities is used to rank the potential arcs joining the entity nodes. The proposed multistrain diffusion model and mutual-strain infection ranking technique have been implemented and tested on widely accepted social network data sets. Experiments show that the MSDM-LP and mutual-strain diffusion ranking technique outperforms state-of-the-art algorithms for neighbor-based ranking.

In this paper, a platform is presented, that given a Stochastic Context-Free Grammar (SCFG), automatically outputs the description of a parser in synthesizable Hardware Description Language (HDL) which can be downloaded in an FPGA (Field Programmable Gate Arrays) board. Although the proposed methodology can be used for various inexact models, the probabilistic model is analyzed in detail and the extension to other inexact schemes is described. Context-Free Grammars (CFG) are augmented with attributes which represent the probability values. Initially, a methodology is proposed based on the fact that the probabilities can be evaluated concurrently with the parsing during the parse table construction by extending the fundamental parsing operation proposed by Chiang & Fu. Using this extended operation, an efficient architecture is presented based on Earley’s parallel algorithm, which given an input string, generates the parse table while evaluating concurrently the probabilities of the generated dotted grammar rules in the table. Based on this architecture, a platform has been implemented that automatically generates the hardware design of the parser given a SCFG. The platform is suitable for embedded systems applications where a natural language interface is required or in pattern recognition tasks. The proposed hardware platform has been tested for various SCFGs and was compared with previously presented hardware parser for SCFGs based on Earley’s parallel algorithm. The hardware generated by the proposed platform is much less complicated than the one of comparison and succeeds a speed-up of one order of magnitude.