基于视频的现实人物运动创作

提出一种根据用户指定的人体运动和观察视角生成真实感视频的方法.首先采集演员进行少数基本运动时的多视角视频数据库,并使用无标记运动捕捉的方法获得任意时刻人体对应的骨骼和3D模型.其次,用户对人体骨架指定运动并设定视角,以此定义目标视频.实验结果验证了文中方法能够利用有限的数据库合成演员在用户指定运动和视角下的真实感视频.     

稀疏差分网络和多监督哈希用于高效图像检索

针对基于深度哈希的图像检索中卷积神经网络（CNN）特征提取效率较低和特征相关性利用不充分的问题,提出一种融合稀疏差分网络和多监督哈希的新方法SDNMSH（sparse difference networks and multi-supervised hashing）,并将其用于高效图像检索。SDNMSH以成对的图像作为训练输入,通过精心设计的稀疏差分卷积神经网络和一个监督哈希函数来指导哈希码学习。稀疏差分卷积神经网络由稀疏差分卷积层和普通卷积层组成。稀疏差分卷积层能够快速提取丰富的特征信息,从而实现整个网络的高效特征提取。同时,为了更加充分地利用语义信息和特征的成对相关性,以促进网络提取的特征信息能够更加有效地转换为具有区分性的哈希码、进而实现SDNMSH的高效图像检索,采用一种多监督哈希（MSH）函数,并为此设计了一个目标函数。在MNIST、CIFAR-10和NUS-WIDE三个广泛使用的数据集上进行了大量的对比实验,实验结果表明,与其他先进的深度哈希方法相比,SDNMSH取得了较好的检索性能。     

新型的数字音频系统—数字盒式磁带（DCC）原理

数字盒式磁带录音机,简称DCC,是一种将现有的盒式磁带与数字声频相结合的产物。它采用了新的有效的精确自适应子带编码技术,简称PASC,使重放的声音达到CD的音质。PASC依据数字声频中两个新的原理:与人耳的听觉特性相适应原理和有效的自适原应理。DCC系统可以与模拟音乐盒带兼容,现有的盒带可以在DCC放音设备上重放。     

强度传输方程和神经网络融合的三维重构算法

针对原有强度传输方程法所恢复的相位精度不够精确的缺点,提出强度传输方程和神经网络融合的三维重构算法. 利用强度传输方程求解出物体不同角度的初始相位,利用神经网络算法进行优化,根据不同角度的最终恢复相位结合乘法技术重构出三维体信息. 该算法具有精度高的特点,可以为三维成像技术的应用提供参考. 对于实验中的示例图像,该算法将强度传输方程得到的相位误差从21.40%降低为5.26%,重构三维物体与模拟真实物体的相关程度为显著相关.     

一种新颖的低层视觉特征到高层语义的映射方法*

通过在低层视觉特征上提取图像的局部颜色和纹理特征,再利用模糊神经网络建立低层视觉特征和高层语义特征之间的映射,实现了图像属性信息的有效传递和可靠检索.最后,为了提高检索效率,把相关反馈引进到图像检索系统中.实验证明,该方法取得了较好的检索查全率和准确率.     

An architectural design for effective information retrieval in semantic web

The current web IR system retrieves relevant information only based on the keywords which is inadequate for that vast amount of data. It provides limited capabilities to capture the concepts of the user needs and the relation between the keywords. These limitations lead to the idea of the user conceptual search which includes concepts and meanings. This study deals with the Semantic Based Information Retrieval System for a semantic web search and presented with an improved algorithm to retrieve the information in a more efficient way.This architecture takes as input a list of plain keywords provided by the user and the query is converted into semantic query. This conversion is carried out with the help of the domain concepts of the pre-existing domain ontologies and a third party thesaurus and discover semantic relationship between them in runtime. The relevant information for the semantic query is retrieved and ranked according to the relevancy with the help of an improved algorithm. The performance analysis shows that the proposed system can improve the accuracy and effectiveness for retrieving relevant web documents compared to the existing systems.     

Ensemble of multiple instance classifiers for image re-ranking

Text-based image retrieval may perform poorly due to the irrelevant and/or incomplete text surrounding the images in the web pages. In such situations, visual content of the images can be leveraged to improve the image ranking performance. In this paper, we look into this problem of image re-ranking and propose a system that automatically constructs multiple candidate “multi-instance bags (MI-bags)”, which are likely to contain relevant images. These automatically constructed bags are then utilized by ensembles of Multiple Instance Learning (MIL) classifiers and the images are re-ranked according to the final classification responses. Our method is unsupervised in the sense that, the only input to the system is the text query itself, without any user feedback or annotation. The experimental results demonstrate that constructing multiple instance bags based on the retrieval order and utilizing ensembles of MIL classifiers greatly enhance the retrieval performance, achieving on par or better results compared to the state-of-the-art.     

Efficient data selection for speech recognition based on prior confidence estimation using speech and monophone models

This paper proposes an efficient speech data selection technique that can identify those data that will be well recognized. Conventional confidence measure techniques can also identify well-recognized speech data. However, those techniques require a lot of computation time for speech recognition processing to estimate confidence scores. Speech data with low confidence should not go through the time-consuming recognition process since they will yield erroneous spoken documents that will eventually be rejected. The proposed technique can select the speech data that will be acceptable for speech recognition applications. It rapidly selects speech data with high prior confidence based on acoustic likelihood values and using only speech and monophone models. Experiments show that the proposed confidence estimation technique is over 50 times faster than the conventional posterior confidence measure while providing equivalent data selection performance for speech recognition and spoken document retrieval.     

结合相关规则和本体加权图的查询扩展

针对现有信息检索系统查询性能的不足,提出了一种结合相关规则和WordNet本体信息的查询扩展方法.该方法借助相关规则挖掘和WordNet本体信息构建加权词语关系图,并根据加权图的结构和权重信息计算扩展词的重要性.查询时,从这个图中取原查询词的最邻近词作为扩展词来源,选取其中权重最大的p个词返回并进行二次检索.在实现算法的基础上,通过Lucene全文检索器进行实验,将所得的结果值F1与其他算法的结果作比较.比较结果表明,该方法比不作扩展的检索有约16.93％的性能提升.