手写体汉字识别问题综论

本文在讨论了人对汉字的认知心理试验研究结果之后,认为人类的汉字识别是经由整字属性及分层结构的多种途径复合而成的。这些途径中的大多数都可以在手写体汉字的机器识别中模仿运用。由此可把汉字的机器识别方法归纳为整模式法和子模式法两个大类,对它们的有利方面和困难方面进行了讨论;对影响其性能的因素进行了分析;对识别率以及它与识别字集大小和后处理等的关系提出了评估的问题。     

激光辐照对冷冻精子活率和活力的影响

分别用7.48J/cm~2和14.96J/cm~2能量密度的氦氖激光辐照23份液氮冷冻前人的精液,冻贮三个月后复温查精子活率和毛细管穿透高度。结果:经7.48J/cm~2剂量的激光辐照后的精子活率和穿透高度明显提高,分别平均提高10.18％和8.32mm,与未受辐照组比较差异显著(P<0.05)经14.96J/cm~2辐照组,活率较未照组降低。实验结果说明适当剂量的激光辐照能够增强精子对冷冻的耐受力,提高精子的复活率和活动力。     

Ranking with uncertain labels and its applications

The techniques for image analysis and classification generally consider the image sample labels fixed and without uncertainties. The rank regression problem studied in this paper is based on the training samples with uncertain labels, which often is the case for the manual estimated image labels. A core ranking model is designed first as the bilinear fusing of multiple candidate kernels. Then, the parameters for feature selection and kernel selection are learned simultaneously by maximum a posteriori for given samples and uncertain labels. The provable convergency Expectation Maximization (EM) method is used for inferring these parameters in an iterative manner. The effectiveness of the proposed algorithm is finally validated by the extensive experiments on age ranking task and human tracking task. The popular FG-NET and the large scale Yamaha aging database are used for the age estimation experiments, and our algorithm outperforms those state-of-the-art algorithms ever reported by other interrelated literatures significantly. The experiment result of human tracking task also validates its advantage over conventional linear regression algorithm. A short version of this paper appeared in ICME07.     

计算机生成兵力中的人工智能技术

开发CGF的难点和重点在于CGF实体行为的生成,尤其是人类智能行为的实现。人类行为建模最重要的三个方面是知识获取、知识表示和决策机制,都与人工智能技术相关。该文主要介绍人类行为描述的知识表示和知识获取问题,讨论了人工智能技术在这两个方面的应用,并对将来的研究工作做了一个简单的展望。     

网络条件下的资本一般及其逻辑推演——利益、规则与标准

通过对资本一般概念的分析，结合资本一般的运动功能论述，认为在经济全球化和网络条件下，资本增值属性必然导致对规则和标准的争夺与控制，标准战是资本一般在新的经济社会形态下发展的必然逻辑结果。     

Human catalase is imported and assembled in peroxisomes of Saccharomyces cerevisiae

To study the conservation of peroxisomal targeting signals, we have determined the intracellular localization of human peroxisomal catalase when expressed in yeast. Using immunofluorescence, differential centrifugation and immunoelectron microscopy, we show that the protein is targeted to the peroxisomes of the heterologous cell and assembled in its active tetrameric form. These data show the conservation of the catalase targeting signal and import specificity between human and yeast peroxisomes.     

手势数据驱动的头部运动合成方法

非手部手势是手语表达中不可缺少的一部分,头部运动的实现并与手势进行协同表达是其重要研究内容。对真人手语表演数据中的手势与头部动作之间的关系进行了深入研究,提取二者的动作特征,利用核典型相关分析方法(kernel canonical correlation analysis,KCCA)建立起手势与头部动作之间的预测关系模型。动画合成结果以及评价实验表明,KCCA方法能更好地刻画手势与头部动作的协调性,实现虚拟人行为动作合成的逼真性。     

近20年塔里木河干流生态环境变化遥感监测分析

针对塔里木河流域综合治理后生态环境的变化状况及植被的恢复状况等热点问题,基于遥感时空变化分析方法从土地利用、地表水体和植被覆盖度的时空变化方面分析了近20年来塔里木河干流生态环境的变化状况。结果表明:(1)塔里木河干流土地利用受到人类活动的影响显著,自2000年以来,塔里木河干流耕地面积增加显著,净增2 229 km~2,草地面积整体下降,萎缩3 157 km~2;(2)综合治理以来,下游河段2009—2018时段水体出现频率明显高于2000—2008时段,水体出现频率提高约15%;(3)2000年以来,塔里木河流域的植被覆盖度持续增加,特别是下游河段植被覆盖度增加明显,覆盖度增加18%。但中、上游流域的植被覆盖度的增加趋势主要贡献来源于耕地,中上游地区耕地的增加仍会对塔里木河流域的生态保护和恢复造成一定的影响。     

3D human pose estimation in motion based on multi-stage regression

3D human pose estimation in motion is a hot research direction in the field of computer vision. However, the performance of the algorithm is affected by the complexity of 3D spatial information, self-occlusion of human body, mapping uncertainty and other problems. In this paper, we propose a 3D human joint localization method based on multi-stage regression depth network and 2D to 3D point mapping algorithm. First of all, we use a single RGB image as the input, through the introduction of heatmap and multi-stage regression to constantly optimize the coordinates of human joint points. Then we input the 2D joint points into the mapping network for calculation, and get the coordinates of 3D human body joint points, and then to complete the 3D human body pose estimation task. The MPJPE of the algorithm in Human3.6 M dataset is 40.7. The evaluation of dataset shows that our method has obvious advantages.     

Safe and intuitive manual guidance of a robot manipulator using adaptive admittance control towards robot agility

Key areas of robot agility include methods that increase capability and flexibility of industrial robots and facilitate robot re-tasking. Manual guidance can achieve robot agility effectively, provided that a safe and smooth interaction is guaranteed when the user exerts an external force on the end effector. We approach this by designing an adaptive admittance law that can adjust its parameters to modify the robot compliance in critical areas of the workspace, such as near and on configuration singularities, joint limits, and workspace limits, for a smooth and safe operation. Experimental validation was done with two tests: a constraint activation test and a 3D shape tracing task. In the first one, we validate the proper response to constraints and in the second one, we compare the proposed approach with different admittance parameter tuning strategies using a drawing task where the user is asked to guide the robot to trace a 3D profile with an accuracy or speed directive and evaluate performance considering path length error and execution time as metrics, and a questionnaire for user perception. Results show that appropriate response to individual and simultaneous activation of the aforementioned constraints for a safe and intuitive manual guidance interaction is achieved and that the proposed parameter tuning strategy has better performance in terms of accuracy, execution time, and subjective evaluation of users.