多尺度高分辨率保持和视角不变的手姿态估计

目前基于彩色图像的手姿态2D关键点热图估计大多数采用卷积姿势机或沙漏网络进行,但这两种网络不能同时满足高分辨率表示保持学习和多尺度特征融合。针对该问题引用了一种多尺度高分辨率保持的网络,该网络采用高低分辨率表示并行设计的结构,并通过融合所有分辨率表示增强各分辨率表示的特征,而且拥有多个阶段提取高质量特征用于2D热图估计。为得到3D手姿态,还使用了全局旋转视角不变的方法将2D热图映射到3D姿态。在三个公开数据集（RHD、STB、Dexter+Object）上分别对2D手姿态估计和3D手姿态估计进行了实验,结果验证了该方法在手姿态估计中的有效性。     

基于聚类的单帧图像超分辨率重建方法

为解决单幅图像的超分辨重建问题,提出一种基于聚类的单帧图像超分辨率重建方法.从高分辨率样本图像中学习一个结构聚类型的高分辨率字典,利用迭代收缩算法优化目标方程,求得高分辨率图像的表示系数,使用学习到的高分辨率字典对低分辨率图像进行重构.实验结果表明,与总变分方法、软切割方法和稀疏表示方法相比,该方法的单帧图像超分辨率重建效果较好.     

三维微小凹图式的加工及其与人神经母细胞瘤细胞的复合

采用紫外光光刻、硅蚀刻及软光刻技术分别制备了阴性光刻胶SU-8和聚乳酸及聚羟基乙酸共聚物(PLGA)微小凹图式.以人神经母细胞瘤细胞SH-SY5Y与三维微结构进行复合,采用扫描电子显微镜、光学显微镜及caicein荧光染色评价所加工图式的结构形态及细胞的生长行为.实验制备出名义直径约100/μm、高纵横比(接近或大于1)并可叠加微通道连接的SU-8及PLGA三维微小凹图式.所加工的图式透明,符合三维细胞生物传感器的光学检测的需要.细胞培养结果表明,该结构尺寸的图式能引导在三维环境神经细胞的排布、神经突起延伸及形态分化.     

基于Monte Carlo模型的微型PLGA给药系统建模及仿真

多腔体的微型可降解高分子聚合物PLGA药物缓释系统是一种新型植入式给药微器件,其载体结构是结合药物释放的要求和高分子聚合物生物降解特性进行设计并利用MEMS工艺制备.为了解微型给药系统实际释药的性能,需要对其进行建模和仿真研究.基于体溶蚀的Monte Carlo溶蚀模型,建立了具有多腔体的微型PLGA给药载体的释药模型,并对腔体结构为圆形的微型给药系统进行了释药过程仿真.仿真结果表明本文建立的微系统释药模型可以较为准确的描述微系统的释药过程,仿真模型对进一步开发微型PLGA给药系统有重要的参考价值.     

单材料几何微结构设计研究综述

增材制造(3D打印)技术的快速发展促进了复杂几何结构在航空航天、辅助医疗、交通运输等工业生产中的应用.在自然界中广泛存在并具有优异物理性能的几何微结构可以通过3D打印的方式进行制备.如何设计几何微结构以达到特定物理性能目标的问题已成为计算机辅助设计、计算机辅助工程、机械工程与材料科学学科交叉的一项研究热点.本文对近年来面向单材料3D打印的几何微结构设计工作进行综述,系统地梳理了几何微结构的主要设计方法.首先,本文从几何约束和物理目标两个角度介绍了几何微结构的设计要求;随后将几何微结构单元的设计方法从优化方法、参数化方法和过程式方法三个类型分别进行了详细阐述与分析;此外,本文也对几何微结构单元在给定形状空间内的合成方法进行了总结.最后,本文对目前几何微结构设计研究中待解决的问题进行了讨论,并展望了可能的发展方向.     

基于吹塑成型方法的微玻璃空腔的设计与制备

根据硼硅酸盐玻璃的内部结构特殊性和热学性质,设计并制备出两种3D微玻璃空腔,主要讲述了3D微玻璃空腔的设计过程和吹塑成型的制备方法。 CORNING Pyrex 7740玻璃是硼硅酸盐玻璃的代表。将硅片进行深硅刻蚀形成深槽,并与7740玻璃进行常压下的阳极键合,形成微空腔;将得到的微空腔放入真空退火炉中进行退火,使玻璃空腔内部空气膨胀,最终形成3D微玻璃空腔。经过实验得到的两种3D微玻璃空腔表明其制备工艺的可行性,将制备出的3D微玻璃空腔运用到导航器件的设计和微结构的封装等方面,具有比较好的发展前景。     

基于CNTs/PDMS介电层的柔性压力传感特性研究

提出一种基于MEMS工艺的柔性压力传感器制备方法.采用MEMS工艺制备柔性压力传感器模板,结合纳米压印技术、射频磁控溅射技术和PDMS软光刻工艺在PDMS柔性基底上制备了具有"V"型阵列微结构的Ag薄膜平行板电极,基于碳纳米管(CNTs)/PDMS聚合物的压电容特性,制备出电容式柔性压力传感器.针对不同尺寸的压力传感器进行对比测试,本文制作的压力传感器的灵敏度能够达到3.98% kPa-1,具有良好的重复性,在智能穿戴和电子皮肤等方面有着广阔的应用前景.     

一种16位分辨率DAC的设计与实现

DAC0832是一款广泛应用的8位分辨率的D/A转换集成芯片。本文通过对DAC0832的研究和实验,提出了一种利用两片DAC0832实现16位分辨率的D/A转换器的新方法,降低了获得高分辨率D/A转换器的成本。文章中给出了两片DAC0832实现16位分辨率的D/A转换器的硬件电路、程序设计及测试数据。高8位数模转换和低8位数模转换的测试结果表明本设计的D/A转换器具有16位的分辨率,且线性度较好,达到16位分辨率D/A转换器的性能指标。应用这种方法,可将DAC0832的分辨率扩展为24位或者32位。     

用于微惯性器件的ICP刻蚀工艺技术

在微惯性器件加工中,ICP深硅刻蚀技术主要用于梳齿结构的释放.工艺试验中的梳齿结构的最细线条尺寸为2μm,刻蚀深度为40μm,刻蚀的深宽比为20∶1,接近刻蚀设备A601E的加工极限.为了提高刻蚀精度,减小根切和底切效应,本文介绍了一种实现微结构刻蚀的ICP分步工艺的新方法,采用不同的刻蚀工艺条件,初始阶段减小底切效应,减小线条损失,刻蚀的中间阶段保证刻蚀速度,刻蚀的最终阶段减小侧向刻蚀,提高结构释放的一致性.同时通过在刻蚀结构的背面生长200 nm厚Al膜对等离子体的吸附作用减小了根切效应,提高了刻蚀的精度和结构释放的一致性.     

用软光刻技术实现微细结构

本文介绍了一种新的微结构制造技术———软光刻技术 ,它提供了一种方便、有效的和低成本的微米、纳米尺寸微结构的制造方法 .本文着重涉及了软光刻的几项关键技术 :弹性印模、再铸模等 ,阐述了模板的制备方法和原理 ,并应用这种模板实现微阀和微毛细管     

基于多模态成像技术监测基因修饰支架中血管生成及修复临界性骨缺损的研究

基因修饰促血管化的支架是促进骨再生的有效方法之一,它可以将目的基因转移到内源性细 胞中实现生长因子原位、持续表达,诱导内源性细胞的增殖、迁移和分化,从而促进骨组织再生。该 文以慢病毒介导基因修饰多孔 PLGA/nHAp 复合支架诱导血管生成为研究对象,采用静电吸附和低温冷冻方法制备基因修饰多孔 PLGA/nHAp 复合支架;以小鼠顶骨临界性骨缺损为模型,利用多模态双光子及光声显微成像技术在体、实时、连续监测,研究骨修复过程中基因修饰多孔支架诱导血管化的动态过程。体外实验结果显示,慢病毒颗粒从支架上的持续释放长达 5 天,缓释出的病毒颗粒可以有效转染 293T 细胞并表达 PDGF-BB 因子。体内实验结果表明,慢病毒介导基因修饰多孔 PLGA/nHAp 复合支架,可以实现 PDGF-BB 因子原位表达,促进体内局部及系统性干细胞等细胞迁移,加快血管诱导生成并提高骨缺损部位骨组织的再生能力。同时,在该研究中,成功使用并比较了多模态双光子及光 声显微成像技术在体、实时、连续监测 3D 骨组织支架内血管形成的动态变化过程,并验证了基因修饰对于提高 3D 打印支架的生物学反应性的作用。该研究为研究不同支架对血管生成作用的监测与鉴定提供了新的技术手段。     

Optimal design of a 3D-printed scaffold using intelligent evolutionary algorithms

Fabrication of three-dimensional structures has gained increasing importance in the bone tissue engineering (BTE) field. Mechanical properties and permeability are two important requirement for BTE scaffolds. The mechanical properties of the scaffolds are highly dependent on the processing parameters. Layer thickness, delay time between spreading each powder layer, and printing orientation are the major factors that determine the porosity and compression strength of the 3D printed scaffold.In this study, the aggregated artificial neural network (AANN) was used to investigate the simultaneous effects of layer thickness, delay time between spreading each layer, and print orientation of porous structures on the compressive strength and porosity of scaffolds. Two optimization methods were applied to obtain the optimal 3D parameter settings for printing tiny porous structures as a real BTE problem. First, particle swarm optimization algorithm was implemented to obtain the optimum topology of the AANN. Then, Pareto front optimization was used to determine the optimal setting parameters for the fabrication of the scaffolds with required compressive strength and porosity. The results indicate the acceptable potential of the evolutionary strategies for the controlling and optimization of the 3DP process as a complicated engineering problem.     

Soft lithography based on photolithography and two-photon polymerization

Over the past decades, soft lithography has greatly facilitated the development of microfluidics due to its simplicity and cost-effectiveness. Besides, numerous fabrication techniques such as multi-layer photolithography, stereolithography and other methods have been developed to fabricate moulds with complex 3D structures nowadays. But these methods are usually not beneficial for microfluidic applications either because of low resolution or sophisticated fabrication procedures. Besides, high-resolution methods such as two-photon lithography, electron-beam lithography, and focused ion beam are often restricted by fabrication speed and total fabricated volume. Nonetheless, the region of interest in typical microfluidic devices is usually very small while the rest of the structure does not require complex 3D fabrication methods. Herein, conventional photolithography and two-photon polymerization are combined for the first time to form a simple hybrid approach in fabricating master moulds for soft lithography. It not only benefits from convenience of photolithography, but also gives rise to complex 3D structures with high resolution based on two-photon polymerization. In this paper, various tests have been conducted to further study its performance, and a passive micromixer has been created as a demonstration for microfluidic applications.     

Bio-MEMS fabricated artificial capillaries for tissue engineering

In this report, we focus on the microfabrication and cell seeding issues of artificial blood capillaries for tissue engineering. Two different fabrication methods (stainless steel electroforming and silicon electroforming) and a number of materials (PC, Polycarbonate and biocompatible material PLGA, poly lactide-co-glycolides) are implemented to build the vascular network. The vascular network is then used as the scaffold to cultivate the bovine endothelial cell (BEC). During the period of cell cultivation, oxygen and nutrient need to be continuously delivered by a circular pressurizing system. In cell culture, encouraging results are obtained through the dynamical seeding of the BEC on the scaffolds. A systematic cell culture process has been developed after repeated experiments. Successful seeding efficiencies are obtained by using the developed systematic cell culture process.     

Internal architecture design and freeform fabrication of tissue replacement structures

Modeling, design and fabrication of tissue scaffolds with intricate architecture, porosity and pore size for desired tissue properties presents a challenge in tissue engineering. This paper will present the details of our development in the design and fabrication of the interior architecture of scaffolds using a novel design approach. The interior architecture design (IAD) approach seeks to generate layered scaffold freeform fabrication tool path without forming complicated 3D CAD scaffold models. This involves: applying the principle of layered manufacturing to determine the scaffold individual layered process planes and layered contours; defining the 2D characteristic patterns of the scaffold building blocks (unit cells) to form the Interior Scaffold Pattern; and the generating the process tool path for freeform fabrication of these scaffolds with the specified interior architecture. Feasibility studies applying the IAD algorithm to example models with multi-interior architecture and the generation of fabrication planning instructions will also be presented.     

Functionally heterogeneous porous scaffold design for tissue engineering

Porous scaffolds with interconnected and continuous pores have recently been considered as one of the most successful tissue engineering strategies. In the literature, it has been concluded that properly interconnected and continuous pores with their spatial distribution could contribute to perform diverse mechanical, biological and chemical functions of a scaffold. Thus, there has been a need for reproducible and fabricatable scaffold design with controllable and functional gradient porosity. Improvements in Additive Manufacturing (AM) processes for tissue engineering and their design methodologies have enabled the development of controlled and interconnected scaffold structures. However homogeneous scaffolds with uniform porosity do not capture the intricate spatial internal micro architecture of the replaced tissue and thus are not capable of capturing the design. In this work, a novel heterogeneous scaffold modeling is proposed for layered-based additive manufacturing processes. First, layers are generated along the optimum build direction considering the heterogeneous micro structure of tissue. Each layer is divided into functional regions based on the spatial homogeneity factor. An area weight based method is developed to generate the spatial porosity function that determines the deposition pattern for the desired gradient porosity. To design a multi-functional scaffold, an optimum deposition angle is determined at each layer by minimizing the heterogeneity along the deposition path. The proposed methodology is implemented and illustrative examples are also provided. The effective porosity is compared between the proposed design and the conventional uniform porous scaffold design. Sample designed structures have also been fabricated with a novel micro-nozzle biomaterial deposition system. The result has shown that the proposed methodology generates scaffolds with functionally gradient porosity.     

An approach to integrating shape and biomedical attributes in vascular models

Computational models have been used widely in tissue engineering research and have proven to be powerful tools for bio-mechanical analysis (i.e., blood flow, growth models, drug delivery, etc). This paper focuses on developing higher-fidelity models for vascular structures and blood vessels that integrate computational shape representations with biomedical properties and features. Previous work in computer-aided vascular modeling comes from two communities. For those in biomedical imaging, the goal of past research has been to develop image understanding techniques for the interpretation of x-ray, magnetic resonance imaging (MRI), or other radiological data. These representations are predominantly discrete shape models that are not tied to physiological properties. The other corpus of existing work comes from those interested in developing physiological models for vascular growth and behavior based on bio-medical attributes. These models usually either have a highly simplified shape representation, or lack one entirely. Further, neither of these representations are suitable for the kind of interactive modeling required by tissue engineering applications.This paper aims to bridge these two approaches and develop a set of mathematical tools and algorithms for feature-based representation and computer-aided modeling of vascular trees for use in computer-aided tissue engineering applications. The paper offers a multi-scale representation based on swept volumes and a feature-based representation that can attribute the geometric representation with information about blood flow, pressure, and other biomedical properties. The paper shows how the resulting representation can be used as part of an overall approach for designing and visualizing vascular scaffolds. As a real-world example, we show how this computational model can be used to develop a tissue scaffold for liver tissue engineering. Such scaffolds may prove useful in a number of biomedical applications, including the growth of replacement tissue grafts and in vitro study of the pharmacological affects of new drugs on tissue cultures.     

BioCell print utilizing patterning with electrostatically injected droplet (PELID) method

The objective of this study is to fabricate three-dimensional cell structures utilizing patterning with electrostatically injected droplet (PELID) method, because it is preferable to perform laboratory experiments with 3D cell structures in tissue engineering and artificial organ. However, it is difficult to fabricate 3D cell structures, because own weight of the cell is above the bonding force between cells. In this paper, we printed Madin?CDarby canine kidney cells and collagen as scaffolds utilizing the PELID method. We investigated growth of printed cells. Number of printed cells was increased day by day. We investigated the fundamental characteristics on patterning collagen. The printed collagen was thick when the time to print was increased. These results indicated that it is possible to fabricate 3D cell structure.     

骨组织支架的计算机辅助设计方法综述

随着先进的计算机辅助设计和增材制造技术的快速发展,使得制造具有复杂几何结构的骨组织支架成为可能。根据骨组织支架功能设计要求,从几何形态的角度出发将其结构分为规则性多孔结构和不规则多孔结构两大类,并综述了骨组织支架的设计方法,特别强调了两种适合增材制造的设计方法,即三周期极小曲面(TPMS)和拓扑优化。针对骨组织支架结构设计面临的技术挑战,展望了骨组织支架设计方法的可能发展趋势。     

Bayesian computer-aided experimental design of heterogeneous scaffolds for tissue engineering

This paper presents a Bayesian methodology for computer-aided experimental design of heterogeneous scaffolds for tissue engineering applications. These heterogeneous scaffolds have spatial distributions of growth factors designed to induce and direct the growth of new tissue as the scaffolds degrade. While early scaffold designs have been essentially homogenous, new solid freeform fabrication (SFF) processes enable the fabrication of more complex, biologically inspired heterogeneous designs with controlled spatial distributions of growth factors and scaffold microstructures. SFF processes dramatically expand the number of design possibilities and significantly increase the experimental burden placed on tissue engineers in terms of time and cost. Therefore, we use a multi-stage Bayesian surrogate modeling methodology (MBSM) to build surrogate models that describe the relationship between the design parameters and the therapeutic response. This methodology is well suited for the early stages of the design process because we do not have accurate models of tissue growth, yet the success of our design depends on understanding the effect of the spatial distribution of growth factors on tissue growth. The MBSM process can guide experimental design more efficiently than traditional factorial methods. Using a simulated computer model of bone tissue regeneration, we demonstrate the advantages of Bayesian versus factorial methods for designing heterogeneous fibrin scaffolds with spatial distributions of growth factors enabled by a new SFF process.