Explore the vast range of titles available.

TP391%Q958.1; 东北虎(Panthera tigris altaica)重识别对于东北虎的行为研究以及保护具有重要意义,针对在自然环境中东北虎难以被重新识别的问题,提出了一种结合指导学习和特征擦除的东北虎重识别方法.设计了一个包括全局流和局部流的双流网络,局部流以分割后的前景图作为输入,使网络更关注前景部分,全局流以原图作为输入,通过局部流指导全局流学习细粒度特征,同时将模型最后全连接层(FC)的各类别权重参数之间的差异作为指导进行特征擦除,以擦除后的特征向量训练网络,从而增强特征向量中更多元素的辨别能力,最后在测试阶段只使用全局流.在野外东北虎重识别(ATRW)数据集上对该方法进行评估,在单摄像机条件下,mAP达到92.1%;在交叉摄像机条件下,mAP达到72.9%,试验结果表明方法有效.

为了满足信息用户更精确的信息检索需求,对基于知识元的文献内容知识层的检索进行了探析。论述了信息检索的粒度问题、知识元的相关概念、知识元检索理论技术、知识元检索的研究成果。知识元检索的理论技术尚未成熟,需要深入研究与探索。

The deep learning technology has shown impressive performance in various vision tasks such as image classification, object detection and semantic segmentation. In particular, recent advances of deep learning techniques bring encouraging performance to fine-grained image classification which aims to distinguish subordinate-level categories, such as bird species or dog breeds. This task is extremely challenging due to high intra-class and low inter-class variance. In this paper, we review four types of deep learning based fine-grained image classification approaches, including the general convolutional neural networks (CNNs), part detection based, ensemble of networks based and visual attention based fine-grained image classification approaches. Besides, the deep learning based semantic segmentation approaches are also covered in this paper. The region proposal based and fully convolutional networks based approaches for semantic segmentation are introduced respectively.

Efficient and scalable cache coherence protocol is crucial to high-performance servers with sharedmemory. The directory-based cache coherence protocol is more desirable than the snooping-based protocol with respect to the scalability. However, even for the former protocol, scaling to a large number of cores is also challenging due to the additional area requirements of the directories. We observed that a significant percentage of the referenced memory blocks were only accessed by a single core（even in parallel applications） which could be considered as private memory blocks. An intuitive motivation from this observation is that memory blocks accessed by a single core do not require coherence maintenance. The issue is to identify the private block and track the change of its access pattern. We propose a novel hardware approach to（1） dynamically identify the shared memory blocks at the cache block level, and（2） bypass the coherence procedure for the private memory blocks. This approach increases the effectiveness of the directory-based approach and therefore improves the system performance. Experimental results showed that, our approach can on an average（1） avoid the coherence tracking of about 54% referenced memory blocks,（2） reduce the coherence overhead by 77%,（3） avoid 8% L2 cache misses, and（4） shorten the execution time of parallel applications by 13%.

Love for a person involves an idea of the other’s particularity, or “irreplaceability”: an idea that is linked with a fine grained attention to, and affection for, very specific features of the other. This attention is largely a consequence of—or manifestation of—my love, rather than its ground. What I see in the other is partly conditioned by an “unconditional” concern for her as an individual. A consideration of the importance of the face and of personal names in my interactions with another may bring into focus crucial aspects of the particularity and unconditionality that are characteristic of love.

Sliding friction and wear behavior of two different grain-size aluminas （average grain size = 0.6 and 2 μm） was studied in physiological saline environment. Relationship between time and wear loss was observed. The wear loss volume increased with increased sliding time and fine-grained alumina exhibits a lower value. In the initial wear-in period, grains pull-out caused by grain-boundary microcracking is the dominant wear mechanism of submicron grained alumina. In the case of alumina of large grain size, intergranular and transgranular fracture is the basic wear mechanism. As the sliding time increases, compaction of wear debris on sliding surface of fine and coarsen alumina determines the wear characteristics.

The surface quality of fine-grained ZrO2 engineering ceramic were researched using 270# diamond wheel both with and without work-piece two-dimension ultrasonic vibration grinding（WTDUVG）. By AFM images, the surface topography and the micro structure of the two-dimensional ultrasonic vibration grinding ceramics were especially analyzed. The experimental results indicate that the surface roughness is related to grinding vibration mode and the material removal mechanism. Surface quality of WTDUVG is superior to that of conventional grinding, and it is easy for two-dimensional ultrasonic vibration grinding that material removal mechanism is ductile mode grinding.

The mechanical properties and microstructure features of the fine-grained heat-affected zone（FGHAZ） of ASTM4130 steel was investigated by optical microscope（OM）,scanning electron microscope（SEM）,transmission electron microscope（TEM）,and welding thermal simulation test.It is found that serious embrittlement occurs in the FGHAZ with an 81.37% decrease of toughness,compared with that of the base metal.Microstructure analysis reveals that the FGHAZ is mainly composed of acicular,equiaxed ferrite,granular ferrite,martensite,and martensite-austenite（M-A） constituent.The FGHAZ embrittlement is mainly induced by granular ferrite because of carbides located at its boundaries and sub-boundaries.Meanwhile,the existence of martensite and M-A constituent,which distribute in a discontinuous network,is also detrimental to the mechanical properties.

对红白山剖面的粒度分析表明,塔里木盆地腹地晚新生代沉积物的粒度曲线中普遍存在一个超细粒组分。其众数粒径分布比较稳定,平均为0.87 μm,含量变化介于0.3%~10%,在不同成因的沉积物中和在剖面上都具有明显的变化规律。对同一岩性段而言,超细粒组分含量在风成砂中最低,在黄土中较低,在河流相沉积中较高,在湖相泥岩中最高。尽管受岩性变化的影响比较显著,但不同成因沉积物的超细粒组分含量在剖面上的变化基本一致,暗示了其长周期变化受同一驱动因子控制。化学风化和/或成壤作用强度可能起着关键作用。红白山剖面超细粒组分含量在2.8 Ma的快速降低指示了化学风化和/或成壤作用强度的显著减弱,进而反映了塔里木盆地腹地干旱化的显著加强。