W. C. Situ

A parameter estimation method for biological systems modelled by ODE/DDE models using spline approximation and differential evolution algorithm

The inverse problem of identifying unknown parameters of known structure dynamical biological systems, which are modelled by ordinary differential equations or delay differential equations, from experimental data is treated in this paper. A two stage approach is adopted: first, combine spline theory and Nonlinear Programming (NLP), the parameter estimation problem is formulated as an optimization problem with only algebraic constraints; then, a new differential evolution (DE) algorithm is proposed to find a feasible solution. The approach is designed to handle problem of realistic size with noisy observation data. Three cases are studied to evaluate the performance of the proposed algorithm: two are based on benchmark models with priori-determined structure and parameters; the other one is a particular biological system with unknown model structure. In the last case, only a set of observation data available and in this case a nominal model is adopted for the identification. All the test systems were successfully identified by using a reasonable amount of experimental data within an acceptable computation time. Experimental evaluation reveals that the proposed method is capable of fast estimation on the unknown parameters with good precision.

Review of holographic-based three-dimensional object recognition techniques [Invited]

With the advancement of computing and optical technologies, it is now possible to capture digital holograms of real-life object scenes. Theoretically, through the analysis of a hologram, the three-dimensional (3D) objects coded on the hologram can be identified. However, being different from an optical image, a hologram is composed of complicated fringes that cannot be analyzed easily with traditional computer vision methods. Over the years, numerous important research investigations have been attempted to provide viable solutions to this problem. The aim of this work is three-fold. First, we provide a quick walkthrough on the overall development of holographic-based 3D object recognition (H3DOR) in the past five decades, from film-based approaches to digital-based innovation. Second, we describe in more detail a number of selected H3DOR methods that are introduced at different timelines, starting from the late sixties and then from the seventies, where viable digital holographic-based 3D recognition methods began to emerge. Finally, we present our work on digital holographic, pose-invariant 3D object recognition that is based on a recently introduced virtual diffraction plane framework. As our method has not been reported elsewhere, we have included some experimental results to demonstrate the feasibility of the approach.

Affine invariant matching of broken boundaries based on simple genetic algorithm and contour reconstruction

Viewpoint independent identification of fragmented object contours can be accomplished by matching them against a collection of known reference models. For the class of near-planar objects, the matching process can be posed as the search for the existence of an affine transform between a pair of contours. Recently, it has been demonstrated that the search process can be accomplished with the integration of a simple genetic algorithm (SGA) and a quality migrant injection (QMI) operation. The performance is superior to prior arts based on the use of SGA alone in terms of success rates and computation speed. The downside of such approach is the need of more computation time for generating quality migrants in the course of evolution. In this paper, we have proposed a solution to overcome this problem. Our method has two major contributions. The first one is a scheme which enables a closed boundary to be extracted from a set of fragmented object points, and represented as a one-dimensional (1-D) sequence. Second, we have applied SGA to determine the similarity between a pair of closed boundaries by searching the existence of three correspondence point pairs in their 1-D sequences. As a result of these two contributions, the proposed method is substantially faster than the SGA-QMI scheme, and also capable of attaining close to 100% success rate in identifying matched contours.

Enhanced affine invariant matching of broken boundaries based on particle swarm optimization and the dynamic migrant principle

Recently particle swarm optimization (PSO) has been successfully applied in identifying contours that are originated from different views of the same object. As compared with similar approaches based on simple genetic algorithms (SGA), the PSO exhibits higher success rates, faster convergence speed and in general more stable performance. Despite these favorable factors, there are scenarios where the failure rates in matching certain contours are prominently higher than its peers, and the overall performance also deteriorates rapidly with decreasing swarm size. These shortcomings could be attributed to the lack of an initial swarm community which has the quality to reach the global solution. In this paper we first propose a solution to overcome this problem by integrating PSO and the static migrant principle (SMP). The latter is analogous to migrant policy in real life, introducing a fixed and continuous influx of foreign candidates to the swarm community to promote the diversity, and hence the exploration power in the population. Evaluations show that method is less sensitive to the swarm size, and exhibits moderate enhancement in the success rates as compared with the use of PSO alone. To further improve the performance, we introduce the dynamic migrant principle (DMP) to adjust the balance between exploration and exploitation throughout the optimization process. With this approach high success rates are attained for all test samples based on a small swarm community. In addition, the incorporation of both versions of the migrant principle does not impose any overhead on the complexity of the matching scheme.

A Graphics Processing Unit Accelerated Genetic Algorithm for Affine Invariant Matching of Broken Contours

Past research works have demonstrated matching of fragmented contours can be effectively accomplished with the integration of genetic algorithms and migrant principle. Despite the success, the computation involved in the evaluation of the fitness function is substantial. To overcome this problem, a new formulation on the fitness evaluation targeted for graphics processing unit (GPU) has been developed and presented in this paper. Experimental results reveal that the proposed solution is capable of reducing the matching time while maintaining high success rates.

Fast generation of hologram from range camera images based on the sub-lines and holographic interpolation

The intensity image and the depth images of a three-dimensional object scene can be captured with a commodity range camera, and converted into a Fresnel hologram. However, for some cameras, the images are subject to radial distortion, and too small to be visible in optical reconstruction. Moreover, the conventional hologram generation process with numerical means is significant. In this paper, we present a fast method to overcome the above-mentioned problems. First, the intensity and the depth images are transformed to reduce the radial distortion. Next, the images are interpolated horizontally, and converted into a sequence of sub-lines. Finally, the sub-lines are swiftly converted into a Fresnel hologram through padding along the vertical direction. The pair of interpolation process effectively increases the size and visibility of the reconstructed image. Although our method can be applied to different kinds of range cameras, we have selected the Swissranger model as a showcase to demonstrate the feasibility of the approach.

Affine invariant matching of broken boundaries in noisy images based on the quality migrant injection genetic algorithm and a successive erosion and distance accumulation scheme

Viewpoint invariant identification of fragmented scene contours can be realized by matching them against a collection of known reference models. For near planar objects, the matching of a pair of contours can be encapsulated as the search for the existence of an affine transform between them. Past research has demonstrated that the search process can be effectively accomplished with the integration of a simple genetic algorithm (SGA) and quality migrant injection (QMI), a method referred to as the quality migrant genetic algorithm (QMGA). Despite the favorable outcome, this method is extremely vulnerable to noise contamination on the image scene. In this paper we provide an explanation on the causes of this problem, and propose a solution known as successive erosion and distance accumulation (SEDA). Experimental evaluation shows that by supplementing the QMGA method with the proposed scheme, higher success rates can be attained in identifying matched contours under moderate amount of noise contamination.

Fast affine invariant shape matching from 3d images based on the distance association map and the genetic algorithm

The decision on whether a pair of closed contours is derived from different views of the same object, a task commonly known as affine invariant matching, can be encapsulated as the search for the existence of an affine transform between them. Past research has demonstrated that such search process can be effectively and swiftly accomplished with the use of genetic algorithms. On this basis, a successful attempt was developed for the heavily broken contour situation. In essence, a distance image and a correspondence map are utilized to recover a closed boundary from a fragmented scene contour. However, the pre-processing task involved in generating the distance image and the correspondence map consumes large amount of computation. This paper proposes a solution to overcome this problem with a fast algorithm, namely labelled chamfer distance transform. In our method, the generation of the distance image and the correspondence map is integrated into a single process which only involves small amount of arithmetic operations. Evaluation reveals that the time taken to match a pair of object shapes is about 10 to 30 times faster than the parent method.

Comparison between the applications of fragment-based and vertex-based GPU approaches in k-means clustering of time series gene expression data

With the emergence of microarray technology, clustering of gene expression data has become an area of immense interest in recent years. However, due to the high dimensionality and complexity of the gene data landscape, the clustering process generally involves enormous amount of arithmetic operations. The problem has been partially alleviated with the K-Means algorithm, which enables high dimension data to be clustered efficiently. Further enhancement on the computation speed is achieved with the use of fragment shader running in a graphic processing unit (GPU) environment. Despite the success, such approach is not optimal as the process is scattered between the CPU and the GPU, causing bottleneck in the data exchange between the two processors, and the underused of the GPU. In this paper, we propose to realize the K-Means clustering algorithm with an integration of the vertex and the fragment shaders, which enables the majority of the clustering process to be implemented within the GPU. Experimental evaluation reflects that the computation efficiency of our proposed method in clustering short time gene expression is around 1.5 to 2 times faster than that attained with the conventional fragment shaders.

Low Complexity Compression of Hologram Sub-Lines

In this paper we propose a low complexity scheme for compressing hologram sub-lines based on Predictive coding. Our method can attain a compression ratio of 16 times with only slight artifacts on the reconstructed images.