Yangfan Wang

Retinal vessel segmentation using multiwavelet kernels and multiscale hierarchical decomposition

We propose a comprehensive method for segmenting the retinal vasculature in fundus camera images. Our method does not require preprocessing and training and can therefore be used directly on different images sets. We enhance the vessels using matched filtering with multiwavelet kernels (MFMK), separating vessels from clutter and bright, localized features. Noise removal and vessel localization are achieved by a multiscale hierarchical decomposition of the normalized enhanced image. We show a necessary condition to achieve the optimal decomposition and derive the associated value of the scale parameter controlling the amount of details captured. Finally, we obtain a binary map of the vasculature by locally adaptive thresholding, generating a threshold surface based on the vessel edge information extracted by the previous processes. We report experimental results on two public retinal data sets, DRIVE and STARE, demonstrating an excellent performance in comparison with retinal vessel segmentation methods reported recently.

A Multiscale Moving Boundary Model Arising in Cancer Invasion

Cancer invasion of tissue is a key aspect of the growth and spread of cancer and is crucial in the process of metastatic spread, i.e., the growth of secondary cancers. Invasion consists in cancer cells secreting various matrix degrading enzymes (MDEs) which destroy the surrounding tissue or extracellular matrix (ECM). Through a combination of proliferation and migration, the cancer cells then actively spread locally into the surrounding tissue. Thus processes occurring at the level of individual cells eventually give rise to processes occurring at the tissue level. In this paper we introduce a new type of multiscale model describing the process of cancer invasion of tissue. Our multiscale model is a two-scale model which focuses on the macroscopic dynamics of the distributions of cancer cells and of the surrounding extracellular matrix, and on the microscale dynamics of the MDEs, produced at the level of the individual cancer cells. These microscale dynamics take place at the interface of the cancer cells...

Continuous finite element schemes for a phase field model in two-layer fluid Bénard–Marangoni convection computations

Abstract In this article, we study a phase field model for a two-layer fluid where the temperature dependence of both the density (buoyancy forces) and the surface tension (Marangoni effects) is considered. The phase field model consisting of a modified Navier–Stokes equation, a Cahn–Hilliard phase field equation and an energy transport equation is derived through an energetic variational procedure. An appropriate variational form and a continuous finite element method are adopted to maintain the underlying energy law to its greatest extent. A few examples for Benard–Marangoni convection in an Acetonitrile and n-Hexane two-layer fluid system heated from above will be computed to justify our phase field model and further show the good performance of our methods. In addition, an interesting experiment will be performed to show the competition between the Marangoni effects and the buoyancy forces.

ROV-based Underwater Vision System for Intelligent Fish Ethology Research

Fish ethology is a prospective discipline for ocean surveys. In this paper, one ROV-based system is established to perform underwater visual tasks with customized optical sensors installed. One image quality enhancement method is first presented in the context of creating underwater imaging models combined with homomorphic filtering and wavelet decomposition. The underwater vision system can further detect and track swimming fish from the resulting images with the strategies developed using curve evolution and particular filtering, in order to obtain a deeper understanding of fish behaviours. The simulation results have shown the excellent performance of the developed scheme, in regard to both robustness and effectiveness.

LMI-Based Approach for Exponential Robust Stability of High-Order Hopfield Neural Networks with Time-Varying Delays

This paper studies the problems of global exponential robust stability of high-order hopfield neural networks with time-varying delays. By employing a new Lyapunov-Krasovskii functional and linear matrix inequality, some criteria of global exponential robust stability for the high-order neural networks are established, which are easily verifiable and have a wider adaptive.

Technical note: an R package for fitting sparse neural networks with application in animal breeding1

Neural networks (NNs) have emerged as a new tool for genomic selection (GS) in animal breeding. However, the properties of NN used in GS for the prediction of phenotypic outcomes are not well characterized due to the problem of over-parameterization of NN and difficulties in using whole-genome marker sets as high-dimensional NN input. In this note, we have developed an R package called snnR that finds an optimal sparse structure of a NN by minimizing the square error subject to a penalty on the L1-norm of the parameters (weights and biases), therefore solving the problem of over-parameterization in NN. We have also tested some models fitted in the snnR package to demonstrate their feasibility and effectiveness to be used in several cases as examples. In comparison of snnR to the R package brnn (the Bayesian regularized single layer NNs), with both using the entries of a genotype matrix or a genomic relationship matrix as inputs, snnR has greatly improved the computational efficiency and the prediction ability for the GS in animal breeding because snnR implements a sparse NN with many hidden layers.

Identification and Characterization of Neuropeptides by Transcriptome and Proteome Analyses in a Bivalve Mollusc Patinopecten yessoensis

Neuropeptides play essential roles in regulation of reproduction and growth in marine molluscs. But their function in marine bivalves – a group of animals of commercial importance – is largely unexplored due to the lack of systematic identification of these molecules. In this study, we sequenced and analyzed the transcriptome of nerve ganglia of Yesso scallop Patinopecten yessoensis, from which 63 neuropeptide genes were identified based on BLAST and de novo prediction approaches, and 31 were confirmed by proteomic analysis using the liquid chromatography-tandem mass spectrometry (LC-MS/MS). Fifty genes encode known neuropeptide precursors, of which 20 commonly exist in bilaterians and 30 are protostome specific. Three neuropeptides that have not yet been reported in bivalves were identified, including calcitonin/DH31, lymnokinin and pleurin. Characterization of glycoprotein hormones, insulin-like peptides, allatostatins, RFamides, and some reproduction, cardioactivity or feeding related neuropeptides reveals scallop neuropeptides have conserved molluscan neuropeptide domains, but some (e.g., GPB5, APGWamide and ELH) are characterized with bivalve-specific features. Thirteen potentially novel neuropeptides were identified, including 10 that may also exist in other protostomes, and 3 (GNamide, LRYamide, and Vamide) that may be scallop specific. In addition, we found neuropeptides potentially related to scallop shell growth and eye functioning. This study represents the first comprehensive identification of neuropeptides in scallop, and would contribute to a complete understanding on the roles of various neuropeptides in endocrine regulation in bivalve molluscs.

Registering Retinal Vessel Images from Local to Global via Multiscale and Multicycle Features

We propose a comprehensive method using multiscale and multicycle features for retinal vessel image registration with a local and global strategy. The multiscale vessel maps generated by multiwavelet kernels and multiscale hierarchical decomposition contain segmentation results at varying image resolutions in different levels of vessel details. Then the multicycle feature composed of various combinations of cycle structures with different numbers of vertices is extracted. The cycle structure consisting of vessel bifurcation points, crossover points of arteries and veins, and the connected vessels can be found by our Angle-based Depth-First Search (ADFS) algorithm. Local initial registration is implemented by the matched Cycle-Vessel feature points and global final registration is completed by the Cycle-Vessel-Bifurcation feature points using similarity transformation. Finally, our Skeleton Alignment Error Measure (SAEM) is calculated for optimal scale and cycle feature selection, yielding the best registration result intelligently. Experimental results show that our method outperforms state-of-the-art methods on retinal vessel image registration using different features in terms of accuracy and robustness.