结合位图切割和区域合并的彩色图像分割

就经典分水岭图像分割算法中存在的过分割问题，提出一种结合位图切割和区域合并的彩色图像分割算法。对原始彩色图像通过空域梯度算子求其梯度图像，并利用位图切割重建梯度图像；对新梯度图像进行分水岭预分割；对预分割图像基于异质性最小原则进行区域合并，并获得最终分割结果。相比于现有的同类方法，该算法引入位图切割，抑制噪声对分割结果的影响，在边缘模糊处分割准确，得到符合人类视觉的较小分割区域数目，同时在运行效率上提高。     

CLOI-NET: Class segmentation of industrial facilities’ point cloud datasets

Shape segmentation from point cloud data is a core step of the digital twinning process for industrial facilities. However, it is also a very labor intensive step, which counteracts the perceived value of the resulting model. The state-of-the-art method for automating cylinder detection can detect cylinders with 62% precision and 70% recall, while other shapes must then be segmented manually and shape segmentation is not achieved. This performance is promising, but it is far from drastically eliminating the manual labor cost. We argue that the use of class segmentation deep learning algorithms has the theoretical potential to perform better in terms of per point accuracy and less manual segmentation time needed. However, such algorithms could not be used so far due to the lack of a pre-trained dataset of laser scanned industrial shapes as well as the lack of appropriate geometric features in order to learn these shapes. In this paper, we tackle both problems in three steps. First, we parse the industrial point cloud through a novel class segmentation solution (CLOI-NET) that consists of an optimized PointNET++ based deep learning network and post-processing algorithms that enforce stronger contextual relationships per point. We then allow the user to choose the optimal manual annotation of a test facility by means of active learning to further improve the results. We achieve the first step by clustering points in meaningful spatial 3D windows based on their location. Then, we apply a class segmentation deep network, and output a probability distribution of all label categories per point and improve the predicted labels by enforcing post-processing rules. We finally optimize the results by finding the optimal amount of data to be used for training experiments. We validate our method on the largest richly annotated dataset of the most important to model industrial shapes (CLOI) and yield 82% average accuracy per point, 95.6% average AUC among all classes and estimated 70% labor hour savings in class segmentation. This proves that it is the first to automatically segment industrial point cloud shapes with no prior knowledge at commercially viable performance and is the foundation for efficient industrial shape modeling in cluttered point clouds.     

基于深度学习的秀丽隐杆线虫显微图像分割方法

利用计算机实现自动、准确的秀丽隐杆线虫(C.elegans)的各项形态学参数分析,至关重要的是从显微图像上分割出线虫体态,但由于显微镜下的图像噪声较多,线虫边缘像素与周围环境相似,而且线虫的体态具有鞭毛和其他附着物需要分离,多方面因素导致设计一个鲁棒性的C.elegans分割算法仍然面临着挑战。针对这些问题,提出了一种基于深度学习的线虫分割方法,通过训练掩模区域卷积神经网络(Mask R-CNN)学习线虫形态特征实现自动分割。首先,通过改进多级特征池化将高级语义特征与低级边缘特征融合,结合大幅度软最大损失(LMSL)损失算法改进损失计算;然后,改进非极大值抑制;最后,引入全连接融合分支等方法对分割结果进行进一步优化。实验结果表明,相比原始的Mask R-CNN,该方法平均精确率(AP)提升了4.3个百分点,平均交并比(mIOU)提升了4个百分点。表明所提出的深度学习分割方法能够有效提高分割准确率,在显微图像中更加精确地分割出线虫体。     

A Topical Formulation of Melatoninergic Compounds Exerts Strong Hypotensive and Neuroprotective Effects in a Rat Model of Hypertensive Glaucoma

Melatonin is of great importance for regulating several eye processes, including pressure homeostasis. Melatonin in combination with agomelatine has been recently reported to reduce intraocular pressure (IOP) with higher efficacy than each compound alone. Here, we used the methylcellulose (MCE) rat model of hypertensive glaucoma, an optic neuropathy characterized by the apoptotic death of retinal ganglion cells (RGCs), to evaluate the hypotensive and neuroprotective efficacy of an eye drop nanomicellar formulation containing melatonin/agomelatine. Eye tissue distribution of melatonin/agomelatine in healthy rats was evaluated by HPLC/MS/MS. In the MCE model, we assessed by tonometry the hypotensive efficacy of melatonin/agomelatine. Neuroprotection was revealed by electroretinography; by levels of inflammatory and apoptotic markers; and by RGC density. The effects of melatonin/agomelatine were compared with those of timolol (a beta blocker with prevalent hypotensive activity) or brimonidine (an alpha 2 adrenergic agonist with potential neuroprotective efficacy), two drugs commonly used to treat glaucoma. Both melatonin and agomelatine penetrate the posterior segment of the eye. In the MCE model, IOP elevation was drastically reduced by melatonin/agomelatine with higher efficacy than that of timolol or brimonidine. Concomitantly, gliosis-related inflammation and the Bax-associated apoptosis were partially prevented, thus leading to RGC survival and recovered retinal dysfunction. We suggest that topical melatoninergic compounds might be beneficial for ocular health.     

自适应天牛须优化与K均值聚类的图像分割算法研究

在机器识别中,图像分割是重要的一个步骤,传统分割手段存在一定缺陷。针对传统K均值聚类分割的初始聚类中心敏感的缺陷进行了优化,利用自适应天牛须优化算法,避免了这一问题。通过实验结果表明,该算法(ABASK)对图像进行分割,既可以保证图像轮廓的分割,同时还可以更多地保留图像细节。     

用修正淬透性系数法对反应堆压力容器锻件RTNDT数值的预估

分析反应堆压力容器（RPV）锻件制造过程中关键化学元素含量及基准无塑性转变温度（RTNDT）数值，通过引入修正淬透性系数，建立了反映RTNDT变化规律的修正淬透性系数函数表达式，并给出了实例验证，结果表明，通过本文建立的修正淬透性系数函数式可以比较精确地预估RPV锻件的RTNDT值。      

基于阶段注意力机制的电力线提取算法

电力系统维护是电力系统稳定运行的重要保障，应用智能算法的无人机电力巡检则为电力系统维护提供便捷。电力线提取是自主电力巡检以及保障飞行器低空飞行安全的关键技术，结合深度学习理论进行电力线提取是电力巡检的重要突破点。本文将深度学习方法用于电力线提取任务，结合电力线图像特点嵌入改进的图像输入策略和注意力模块，提出一种基于阶段注意力机制的电力线提取模型（SA-Unet）。本文提出的SA-Unet模型编码阶段采用阶段输入融合策略（Stage input fusion strategy， SIFS），充分利用图像的多尺度信息减少空间位置信息丢失。解码阶段通过嵌入阶段注意力模块（Stage attention module，SAM）聚焦电力线特征，从大量信息中快速筛选出高价值信息。实验结果表明，该方法在复杂背景的多场景中具有良好的性能。     

Driving behavior recognition and prediction based on Bayesian model

Since the existing intelligent driving systems are lack of efficiency and accuracy when processing huge number of driving data,a brand new approach of processing driving data was developed to identify and predicate human driving behavior based on Bayesian model.The approach was proposed to take two steps to deduce the specific driving behavior from driving data correspondingly without any supervision,the first step being using Bayesian model segmentation algorithm to divide driving data that inertial sensor collected into near-linear segments with the help of Bayesian model segmentation algorithm,and the second step being using extended LDA model to aggregate those linear segments into specific driving behavior (such as braking,turning,acceleration and coasting).Both offline and online experiments are conducted to verify this approach and it turns out that approach has higher efficiency and recognition accuracy when dealing with numerous driving data.     

基于线特征及迭代最近点算法的地基建筑物点云自动配准方法

为克服迭代最近点(ICP)算法易陷入局部最优的缺陷，提出一种基于线特征及ICP算法的地基建筑物点云自动配准方法。首先，基于法向一致性进行建筑物点云平面分割；接着，采用alpha-shape算法进行点簇轮廓线提取，并拆分和拟合处理得到特征线段；然后，以线对作为配准基元，以线对夹角和距离作为相似性测度进行同名特征匹配，实现建筑物点云的粗配准；最后，以粗配准结果为初值，进一步采用ICP算法完成点云精确配准。利用两组部分重叠的建筑物点云进行配准实验，实验结果表明，采用由粗到精的配准方法能有效改善ICP算法对初值依赖的问题，实现具有部分重叠的建筑物点云的有效配准。     

基于自监督学习的病理图像层次分割

在苏木精-伊红(HE)染色病理图像中，细胞染色分布的不均匀和各类组织形态的多样性给病理图像的自动分割带来极大挑战。为解决该问题，提出了一种基于自监督学习的病理图像三步层次分割方法，对病理图像中各类组织进行由粗略到精细的全自动逐层分割。首先，根据互信息的计算结果在RGB色彩空间中进行特征选择；其次，采用K-means聚类将图像初步分割为各类组织结构的色彩稳定区域与模糊区域；然后，以色彩稳定区域为训练集采用朴素贝叶斯分类对模糊区域进行进一步分割，得到完整的细胞核、细胞质和胞外间隙这三类组织结构；最后，对细胞核部分进行结合形状和色彩强度的混合分水岭分割得到细胞核间的精确边界，进而量化计算细胞核个数、核占比、核质比等指标。对脑膜瘤HE染色病理图像的分割实验结果表明，所提方法对于染色和细胞形态差异保持较高的鲁棒性，各类组织区域分割误差在5%以内，在细胞核分割精度的对比实验中平均正确率在96%以上，满足临床自动图像分析的要求，其量化结果可以为定量病理分析提供依据。