乐谱图像中的音符识别方法

面向多声部乐谱,实现基于结构模式的音符识别方法。在音符基元抽取阶段,提出基于游程分析的符干、符头、符梁3类基元抽取算法,具有较好的抗相交和抗粘连干扰能力。在音符结构分析阶段,采用“作用场”描述音符基元关系,将音符分为6类子结构,可缩小基元搜索范围。在此基础上细分音符结构,遵循关键子结构优先定位原则实现音符基元重组,可降低分析复杂度且具有良好的基元冗余排错能力。实验结果表明,该方法能快速准确识别多声部乐谱中的音符,在音符排列密集、结构复杂时适应能力较强。     

全局结构引导的人造物体参数化基元检测

点云是表达三维数据的常见形式,点云数据提取出的几何基元能够帮助人们快速地理解并处理场景信息,也方便后续其他任务的开展.为了更好地利用人造物体中普遍存在的全局结构关系,增强基元检测过程中全局结构的正向引导,提出参数化基元检测网络——RelationNet,包括2个子模块.首先,为了更好地编码三维点与其所在基元的结构关系,通过空间偏移预测模块预测三维点所在基元中心的偏移向量,提升点对其所在基元的位置感知能力,为后续分割任务提供更多的特征依据;其次,人造物体的基元与基元之间常常具有如平行、垂直、轴对齐等结构关系,为了更好地利用这些关系实现对几何基元检测结果的改进,还包含全局结构关系提取模块,利用基元拟合后获得的参数判断各个基元之间的结构关系,并通过设置相应的损失函数对提取到的结果进行引导监督.在大型ABC数据集与基元监督拟合(SPFN), ParseNet等主流算法进行对比的实验结果表明, RelationNet在基元分割和基元分类任务上的MIoU分别达到85.32%和90.10%,与当前先进方法相比有明显的效果提升.     

联机手写中文速记符的自动识别

手写中文速记符的自动识别是文字类识别中的一个比较特殊的课题.实现了一个联机手写中 文速记符识别系统HCSRS(handwritten Chinese shorthand recognition system),并给出了 对速记符中的基本音符、独立略符、连笔符的识别实验结果.该识别系统采用了以声符、韵 符为基元的结构识别策略.为了实现基元的有效切分,提出了一种基于切分-分析-交互结构 及其相应知识库的自调整切分算法STSA(self-tuning segmentation algorithm),从而大大 改     

基于模板匹配和神经网络的乐谱图象符头基元检测与识别

符头基元的检测与识别是光学乐谱识别技术的核心与关键之一.本文提出一种改进的模板匹配结合神经网络仲裁的方法,对乐谱图象中的符头基元进行检测和识别.在模板匹配阶段,利用乐谱的先验知识对符头的搜索范围进行有效限定,减少了匹配过程中搜索的"盲目性".匹配完成后,提取图象中的垂直直线,对相互连接的候选符头和垂直直线进行编码,通过BP神经网对编码样本的训练,实现音符构成规则的智能学习,最后由神经网络完成符头基元的仲裁确认.实验数据表明,与原有的模板匹配方法比较,该方法的整体识别性能体现出了明显的优势,具有较高的实用价值.     

多声部乐谱音符基元抽取方法

面向复杂版面的多声部乐谱,提出了符干、符头和符梁三类音符基元的抽取方法。提出基于垂直游程编码的粗提取、基于水平游程编码的精检测的符干抽取方法,能够有效克服基元间密集相交、粘连的干扰;在音符先验知识引导下,提出先分割、后特征检测的符头抽取方法,解决了粘连符头的切分难题;提出一种基于块状体分割和特征检测的符梁抽取方法,避开了传统的直线抽取方法所无法处理的符梁粘连难题。实验结果表明,该方法能够有效抽取复杂环境下多声部乐谱的音符基元。     

基于局部结构直方图的图像检索

在图像检索领域中,为更好地利用Julesz基元概念来描述图像内容,提出一种新的图像特征描述方法——局部结构直方图。将彩色图像从RGB空间转换到HSV颜色空间,分别提取边缘方向和颜色信息。定义5种基元类型进行局部结构检测,并利用直方图描述图像特征。实验结果表明,与LBP直方图和Gabor滤波器方法相比,该方法的检索性能较优。     

基于生成式模型的视频基元追踪学习

自然场景视频中含有各种类别的视频基元(video primitives),它们构成了整个高维视频块(video bricks)空间,具有不同的结构维度及复杂度,由空间表观与运动共同描述.视频基元主要有两类:结构视频基元与纹理视频基元.文中使用一个通用生成式模型对两类视频基元进行统一概率建模,每个视频基元的表达能力由其对应的信息增益来度量.利用该度量进行视频基元追踪学习,最终建立一个完整的视频基元集.实验结果显示了文中方法在视频内容表示方面的有效性.     

线形三维模型加载方法研究

分析了线状地物建模和三维平台软件加载线形三维模型中存在的问题,提出了一种以小单元组合的方式参数化定制线状地物,加载线形三维模型的方法。该方法以线状地物的空间坐标作为参数,根据需要构建每种线状地物的三维模型小单元,将其作为基元,以基元的空间位置、姿态和缩放比例为关键参数,给出关键参数的计算方法,按照关键参数将基元组合生成线形三维模型加载到三维平台软件中。该方法参数化定制线状地物,简化了三维线状地物建模的工作量,灵活简便地生成三维场景。     

基于PCA的脱机手写汉字的统计模型及其应用

本文通过分析传统汉字的结构模型所具有的优缺点,提出了建立脱机手写汉字统计模型的理论框架;并利用PCA技术发现大量数据规律性的能力,提出了一种基于PCA技术的脱机手写汉字的统计模型.与传统的结构模型相比,该模型避免了目前还无法解决的准确抽取结构基元的困难,通过以容易抽取的可重构的统计特征作为统计基元,并通过对统计基元变化的整体描述或者说对统计基元相互之间关系的描述,较好地建立了脱机手写汉字的统计模型.根据该模型得到的一些实验结果充分说明了其描述脱机手写汉字的有效性.     

基于层次基元结构的优化匹配算法

匹配是计算机视觉中的一个重要而又必不可少的步骤.本文针对运动图象的对应问题,给出了一个新的基于层次基元结构的匹配算法.整个算法分成两步,第一步是层次基元结构的构建.在使用~2G算子获取零交点后,基于一定规则进行聚类,可以构建零交点、零交轮廓线片段、零交轮廓线这样的层次基元结构.第二步是基于相似度准则的由粗至精的匹配过程.本文采用松驰迭代与约束树搜索相结合的策略,获取优化的匹配结果.对真实图象的实验表明,该算法是有效的.     

Emotion-based music recommendation by affinity discovery from film music

With the growth of digital music, the development of music recommendation is helpful for users to pick desirable music pieces from a huge repository of music. The existing music recommendation approaches are based on a user’s preference on music. However, sometimes, it might better meet users’ requirement to recommend music pieces according to emotions. In this paper, we propose a novel framework for emotion-based music recommendation. The core of the recommendation framework is the construction of the music emotion model by affinity discovery from film music, which plays an important role in conveying emotions in film. We investigate the music feature extraction and propose the Music Affinity Graph and Music Affinity Graph-Plus algorithms for the construction of music emotion model. Experimental result shows the proposed emotion-based music recommendation achieves 85% accuracy in average.     

A music notation construction engine for optical music recognition

Optical music recognition (OMR) systems are used to convert music scanned from paper into a format suitable for playing or editing on a computer. These systems generally have two phases: recognizing the graphical symbols (such as note‐heads and lines) and determining the musical meaning and relationships of the symbols (such as the pitch and rhythm of the notes). In this paper we explore the second phase and give a two‐step approach that admits an economical representation of the parsing rules for the system. The approach is flexible and allows the system to be extended to new notations with little effort—the current system can parse common music notation, Sacred Harp notation and plainsong. It is based on a string grammar and a customizable graph that specifies relationships between musical objects. We observe that this graph can be related to printing as well as recognizing music notation, bringing the opportunity for cross‐fertilization between the two areas of research. Copyright © 2003 John Wiley & Sons, Ltd.     

Learning to embed music and metadata for context-aware music recommendation

Contextual factors greatly influence users’ musical preferences, so they are beneficial remarkably to music recommendation and retrieval tasks. However, it still needs to be studied how to obtain and utilize the contextual information. In this paper, we propose a context-aware music recommendation approach, which can recommend music pieces appropriate for users’ contextual preferences for music. In analogy to matrix factorization methods for collaborative filtering, the proposed approach does not require music pieces to be represented by features ahead, but it can learn the representations from users’ historical listening records. Specifically, the proposed approach first learns music pieces’ embeddings (feature vectors in low-dimension continuous space) from music listening records and corresponding metadata. Then it infers and models users’ global and contextual preferences for music from their listening records with the learned embeddings. Finally, it recommends appropriate music pieces according to the target user’s preferences to satisfy her/his real-time requirements. Experimental evaluations on a real-world dataset show that the proposed approach outperforms baseline methods in terms of precision, recall, F1 score, and hitrate. Especially, our approach has better performance on sparse datasets.     

Unlocking digital music

So why is all this happening, and why now? Surely DRM's interminablebad press and poor consumer perception could not have led tothis move towards the abandonment of DRM in favour of nakeddigital content. Or could it?     

Scheduling a music rehearsal problem with unequal music piece length

This paper studies a music rehearsal scheduling where music pieces require different sets of players and different rehearsal durations. The rehearsal is arranged in multiple days. The players need to show up only on the day that the pieces they play are scheduled. They must show up before the first piece they involve starts and leave after the last one ends. The objectives are to assign music pieces to the rehearsal days to minimize the total number of days that all players have to show up and sequence the music pieces within each day to minimize the total waiting time of the players. We propose a 2-stage methodology to schedule music pieces, which is a combination of a cell formation technique and an integer-programming model. From 77 test problems, the solutions from the proposed method are as good as the optimal solutions from MIP models and are better than the best founds in all the test problems. The computational time is also significantly less.     

The effect of music streaming services on music piracy among college students

Previous studies examining the intention and level of music piracy among young people have postulated that collective attitudes, optimistic biases toward risk, and beliefs about copyright laws are the key factors involved in their decisions. We extend the analysis by studying the impact of music streaming services. Music streaming is an alternative business model in which, for a small fee, the consumer has access to a large set of songs without downloading them onto their devices. Results from a Logit model show that college students who are frequent users of music streaming are also more likely to download music illegally. A plausible explanation is that those engaged in music streaming are also heavy users of computer technology, software downloading, and digital sharing – factors that facilitate the conditions for music piracy. Demographics, collective attitudes, and beliefs about risk and rewards continue to play a key role in explaining music piracy, but their relevance is slightly reduced once controlled for music streaming usage.