Automatic breast tumor detection in ABVS images based on convolutional neural network and superpixel patterns

Breast cancer is one of the most common female malignancies, as well as the second leading cause of mortality for women. Early detection and treatment can dramatically decrease the mortality rate. Recently, automated breast volume scanner (ABVS) has become one of the most frequently used diagnose methods for breast tumor screening because of its operator-independent and reproducible advantages. However, it is a challenging job to obtain the tumors’ accurate locations and shapes by reviewing hundreds of ABVS slices. In this paper, a novel computer-aided detection (CADe) system is developed to reduce clinicians’ reading time and improve the efficiency. The CADe system mainly contains three parts: tumor candidate acquisition, false-positive reduction and tumor segmentation. Firstly, a local phase-based approach is built to obtain breast tumor candidates for further recognition. Subsequently, a convolutional neural network (CNN) is applied to reduce false positives (FPs). The introduction of CNN can help to avoid complicated feature extraction as well as elevate the accuracy and efficiency. Finally, superpixel-based segmentation is used to outline the breast tumor. Here, superpixel-based local binary pattern (SLBP) is proposed to assist the segmentation, which improves the performance. The methods were evaluated on a clinical ABVS dataset whose abnormal cases were manually labeled by an experienced radiologist. The experiment results were mainly composed of two parts. At the FP reduction stage, the proposed CNN achieved 100% and 78.12% sensitivity with FPs/case of 2.16 and 0. At the segmentation stage, our SLBP obtained 82.34% true positive, 15.79% false positive and 83.59% Dice similarity. In summary, the proposed CADe system demonstrated promising potential to detect and outline breast tumors in ABVS images.

     

高精度多环片自动装配控制技术

为了实现对微小环片零件的自动化装配,搭建了自动装配系统.通过4根直线导向轴与4个直线轴承来提高系统的导向精度和刚度.采用直线导轨进行各装配作业模块之间的切换,保证了微小环片零件的自动装配与取出.在环片的装配方向上,螺旋升降机和光栅尺实现环片的位置精度控制.在Lab VIEW编程环境中,采用分层软件架构和模块化控制思想,避免了不必要的数据循环检测与丢失,能够达到环片组件的装配精度要求.控制系统分为系统初始化模块、参数设置模块、装配模块和取出模块,自动装配系统通过各个模块间的相互交流配合完成装配任务.采用本文中自动装配系统装配环片的实验结果表明,环片零件装配的最大位置误差为26μm,垂直度误差为17μm,平均装配时间为75 s/片,可满足环片组件所需的精度要求.     

因素空间理论下基点分类算法研究

目前，基于因素空间理论的背景基提取算法计算过程复杂，初始化必须依赖各因素极值，基点数量提取冗余等原因，未能在应用中取得很好效果。为此，结合内点判别法和知识可继承、可扩展的思想，提出一种计算简单、初始化独立、基点数量小的改进的背景基提取算法。然后，利用改进的背景基提取算法构造出一种全新的数据分类算法-基点分类算法，基点分类算法以提取每一类样本的背景基为预测模型，再通过新定义的λ-背景基，优化预测模型。数值实验表明:基点分类算法原理简单、构造难度小、分类模型泛化能力强，预测能力准确率高，同时严格的模型限定区域又能为识别新类别提供新方法。     

基于密集神经网络的灰度图像着色算法

针对在灰度图像着色领域中，传统算法信息提取率不高、着色效果不理想的问题，提出了基于密集神经网络的灰度图像着色算法，以实现改善着色效果，让人眼更好地观察图片信息的目的。利用密集神经网络的信息提取高效性，构建并训练了一个端到端的深度学习模型，对图像中的各类信息及特征进行提取。训练网络时与原图像进行对比，以逐渐减小网络输出结果的信息、分类等各类型的损失。训练完成后，只需向网络输入一张灰度图片，即可生成一张颜色饱满、鲜明逼真的彩色图片。实验结果表明，引入密集网络后，可有效改善着色过程中的漏色、细节信息损失、对比度低等问题，所提算法着色效果较基于VGG网络及U-Net、双流网络结构、残差网络（ResNet）等性能优异的先进着色算法而言取得了显著的改进。     

Autolysis and the endogenous proteinases characterised in beardless barb (Anematichthys apogon) muscle

Beardless barb is a common fish species used in fermentation of fish paste Ka-pi-plaa. Autolytic profile of beardless barb muscle showed the maximum autolysis was at 50 °C, at both acidic and alkaline pH values. With augmentation concentration of NaCl, autolytic activity slightly decreased. Endogenous proteinases isolated from fish muscle in crude extract forms were also characterised. The acidic proteinases had optimum activity at pH 3.0 and 50°C, and they showed higher proteolytic activity than the alkaline proteinases which were optimally active at pH 9.0 and 50 °C. Proteinases in peak at pH 3.0 were inhibited by pepstatin A, but those in peak at pH 9.0 were highly inhibited by PMSF, TLCK and soybean trypsin inhibitor, suggesting that both aspartic and serine proteinases were existed in beardless barb muscle. The proteinases were stable in pH range of 2.0-5.0 but unstable at the temperatures higher than 40 °C. NaCl suppressed the proteolytic activity, ATP activated the proteinase activity, while CaCl₂, MgCl₂ and CoCl₂ exhibited no influence on the activity. The results implied that cathepsin D is the predominant proteinase responsible for autolysis in beardless barb. The findings were useful to improve the processing and qualities of Ka-pi-plaa product using beardless barb as raw material.     

冷却塔免费供冷在高端酒店项目的应用及探讨

本文分析了在过渡季节如何利用冷却塔为国际高端酒店提供免费冷源,以及过渡季节通过冷却塔为酒店供冷的空调效果与经济效益,并通过具体实施项目案例证明了利用冷却塔提供免费冷源的可行性、经济性及对于节能环保的意义.     

A first-principles investigation on mechanical and metallic properties of titanium carbides under pressure

The titanium carbides are potential candidates to achieve both high hardness and refractory property. We carried out a structural search for titanium carbides at three pressures of 0 GPa, 30 GPa and 50 GPa. A phase diagram of the Ti-C system at 0 K was obtained by elucidating formation enthalpies as a function of compositions, and their mechanical and metallic properties of titanium carbides were investigated systematically. We also discussed the relation of titanium concentration to the both mechanical and metallic properties of titanium carbides. It has been found that the average valence electron density and tractility improved at higher concentrations of titanium, while the degree of covalent bonding directionality decreased. To this effect, the hardness of titanium carbide decreases as the content of titanium increases. Our results indicated that the titanium content significantly affected the metallic properties of the Ti-C system.     

简体接管边缘应力区局部减薄处应力特征

参考《压力容器定期检验规则》第40条,采用有限元方法计算了筒体接管边缘应力区局部减薄处的应力分布,得到了局部减薄处凹坑的尺寸与位置对应力的影响趋势,讨论了应力集中系数随尺寸变化的规律,为筒体接管边缘应力区局部减薄处的强度评定提供了一定的计算依据。     

ZnO基薄膜晶体管的研究

ZnO是一种宽带隙的光电半导体材料,能应用于很多领域,如可用在压敏变阻器、声表面波器件、气敏元件、紫外光探测等。ZnO也可以作为有源层应用于薄膜晶体管(TFT)中。ZnO基薄膜晶体管具有以下突出优势:对于可见光部分平均具有80%以上的透射率,迁移率可以高达36cm2/V·s,开/关电流比大于106,可在较低温度(甚至室温)下制备。基于这些优点,ZnOTFT具有取代有源矩阵液晶显示器中常规a-SiTFT的趋势。同时对ZnOTFT的研究也推动了透明电子学的发展。本文阐述了ZnOTFT优越的电学性能,指出了其目前尚存在的不足,并对其发展前景进行了展望。