Hongwei Zheng

Asynchronous therapy restores motor control by rewiring of the rat corticospinal tract after stroke

Improving stroke recovery by timing treatment Patients recovering from strokes often fight a long uphill battle, with mixed results. Studying the effect of physical training on regeneration from damaged nerves in a model of stroke in rats, Wahl et al. show that timing matters. First, the researchers gave the rats a stroke, which damaged their ability to reach for food pellets with their forelimbs. The researchers then gave them physical training and treated them with an antibody to encourage neural regeneration. The rats improved more when the researchers waited until after the antibody treatment to start the training. Damaged circuits, it seems, need a little time to regrow before being called into action. Science, this issue p. 1250 A rat model of stroke shows that the rebuilding of spinal circuits in response to training is time-sensitive. The brain exhibits limited capacity for spontaneous restoration of lost motor functions after stroke. Rehabilitation is the prevailing clinical approach to augment functional recovery, but the scientific basis is poorly understood. Here, we show nearly full recovery of skilled forelimb functions in rats with large strokes when a growth-promoting immunotherapy against a neurite growth–inhibitory protein was applied to boost the sprouting of new fibers, before stabilizing the newly formed circuits by intensive training. In contrast, early high-intensity training during the growth phase destroyed the effect and led to aberrant fiber patterns. Pharmacogenetic experiments identified a subset of corticospinal fibers originating in the intact half of the forebrain, side-switching in the spinal cord to newly innervate the impaired limb and restore skilled motor function.

Double regularized bayesian estimation for blur identification in video sequences

Blind blur identification in video sequences becomes more important. This paper presents a new method for identifying parameters of different blur kernels and image restoration in a weighted double regularized Bayesian learning approach. A proposed prior solution space includes dominant blur point spread functions as prior candidates for Bayesian estimation. The double cost functions are adjusted in a new alternating minimization approach which successfully computes the convergence for a number of parameters. The discussion of choosing regularization parameters for both image and blur function is also presented. The algorithm is robust in that it can handle images that are formed in variational environments with different types of blur. Numerical tests show that the proposed algorithm works effectively and efficiently in practical applications.

Highly-Automatic MI based multiple 2d/3d image registration using self-initialized geodesic feature correspondences

Intensity based registration methods, such as the mutual information (MI), do not commonly consider the spatial geometric information and the initial correspondences are uncertainty In this paper, we present a novel approach for achieving highly-automatic 2D/3D image registration integrating the advantages from both entropy MI and spatial geometric features correspondence methods Inspired by the scale space theory, we project the surfaces on a 3D model to 2D normal image spaces provided that it can extract both local geodesic feature descriptors and global spatial information for estimating initial correspondences for image-to-image and image-to-model registration The multiple 2D/3D image registration can then be further refined using MI The maximization of MI is effectively achieved using global stochastic optimization To verify the feasibility, we have registered various artistic 3D models with different structures and textures The high-quality results show that the proposed approach is highly-automatic and reliable.

Shot retrieval based on fuzzy evolutionary aiNet and hybrid features

As the multimedia data increasing exponentially, how to get the video data we need efficiently become so important and urgent. In this paper, a novel method for shot retrieval is proposed, which is based on fuzzy evolutionary aiNet and hybrid features. To begin with, the fuzzy evolutionary aiNet algorithm proposed in this paper is utilized to extract key-frames in a video sequence. Meanwhile, to represent a key-frame, hybrid features of color feature, texture feature and spatial structure feature are extracted. Then, the features of key-frames in the same shot are taken as an ensemble and mapped to high dimension space by non-linear mapping, and the result obeys Gaussian distribution. Finally, shot similarity is measured by the probabilistic distance between distributions of the key-frame feature ensembles for two shots, and similar shots are retrieved effectively by using this method. Experimental results show the validity of this proposed method.

Complex 3D shape recovery using hybrid geometric shape features in a hierarchical shape segmentation approach

We present a novel and reliable approach for complex object acquisition and surface registration using hybrid geometric shape features in a hierarchical 3D shape approximation and segmentation approach. First, instead of relying on one type of scanned data, we propose to use hybrid data provided that it can support both global and local geometric shape features. The scanned low-resolution global data supplies the global shape prior for registering the high-resolution local surface patches. Local surfaces can thus be optimally registered requiring less overlap and reducing uncertainty. Second, we cannot directly register huge volumes of data simultaneously due to the memory bottlenecks. We segment the global low-resolution model into several meaningful sub-shapes extending a hierarchical algorithm. The local surfaces can be registered on the sub-shapes respectively and all sub-shapes can be merged and rendered after registration. To verify the reliability of the approach, various 3D models have been acquired. The experiments show compelling results by reconstructing very detailed models of complex objects. The approach can be applied to practical 3D modeling applications.

Image statistics and local spatial conditions for nonstationary blurred image reconstruction

Deblurring is important in many visual systems. This paper presents a novel approach for nonstationary blurred image reconstruction with ringing reduction in a variational Bayesian learning and regularization framework. Our approach makes effective use of the image statistical prior and image local spatial conditions through the whole learning scheme. A nature image statistics based marginal prior distribution is used not only for blur kernel estimation but also for image reconstruction. For an ill-posed blur estimation problem, variational Bayesian ensemble learning can achieve a tractable posterior using an image statistic prior which is translation and scale-invariant. During the deblurring, nonstationary blurry images have stronger ringing effects. We thus propose an iterative reweighted regularization function based on the use of an image statistical prior and image local spatial conditions for perceptual image deblurring.

Introducing dynamic prior knowledge to partially-blurred image restoration

The paper presents an unsupervised method for partially-blurred image restoration without influencing unblurred regions or objects. Maximum a posteriori estimation of parameters in Bayesian regularization is equal to minimizing energy of a dataset for a given number of classes. To estimate the point spread function (PSF), a parametric model space is introduced to reduce the searching uncertainty for PSF model selection. Simultaneously, PSF self-initializing does not rely on supervision or thresholds. In the image domain, a gradient map as a priori knowledge is derived not only for dynamically choosing nonlinear diffusion operators but also for segregating blurred and unblurred regions via an extended graph-theoretic method. The cost functions with respect to the image and the PSF are alternately minimized in a convex manner. The algorithm is robust in that it can handle images that are formed in variational environments with different blur and stronger noise.

Extended mumford-shah regularization in bayesian estimation for blind image deconvolution and segmentation

We present an extended Mumford-Shah regularization for blind image deconvolution and segmentation in the context of Bayesian estimation for blurred, noisy images or video sequences. The Mumford-Shah functional is extended to have cost terms for the estimation of blur kernels via a newly introduced prior solution space. This functional is minimized using Γ-convergence approximation in an embedded alternating minimization within Neumann conditions. Accurate blur identification is the basis of edge-preserving image restoration in the extended Mumford-Shah regularization. One output of the finite set of curves and object boundaries are grouped and partitioned via a graph theoretical approach for the segmentation of blurred objects. The chosen regularization parameters using the L-curve method is presented. Numerical experiments show that the proposed algorithm is efficiency and robust in that it can handle images that are formed in different environments with different types and amounts of blur and noise.

An Edge-Driven Total Variation Approach to Image Deblurring and Denoising

Traditional nonlinear filtering techniques are observed in underutilization of blur identification techniques, and vice versa. To improve blind image restoration, a designed edge-driven nonlinear diffusion operator and a point spread function (PSF) learning term are integrated to total variation regularization. The cost functions are minimized iteratively in an alternate minimization with respect to the estimation of images and PSFs under these conditions. Numerical experiments show that the proposed algorithm is efficient and robust in that it can handle images that are formed in different environments with different types and amounts of blur and noise

Discrete Regularization for Perceptual Image Segmentation via Semi-Supervised Learning and Optimal Control

In this paper, we present a regularization approach on discrete graph spaces for perceptual image segmentation via semi-supervised learning. In this approach, first, a spectral clustering method is embedded and extended into regularization on discrete graph spaces. In consequence, the spectral graph clustering is optimized and smoothed by integrating top-down and bottom-up processes via semi-supervised learning. Second, a designed nonlinear diffusion filter is used to maintain semi-supervised learning, labeling and differences between foreground or background regions. Furthermore, the spectral segmentation is penalized and adjusted using labeling prior and optimal window-based affinity functions in a regularization framework on discrete graph spaces. Experiments show that the algorithm achieves perceptual and optimal image segmentation. The algorithm is robust in that it can handle images that are formed in variational environments.