基于Leap Motion的动态手势识别方法

伴随虚拟现实（Virtual Reality,VR）技术的发展,以及人们对人机交互性能和体验感的要求提高,手势识别作为影响虚拟现实中交互操作的重要技术之一,其精确度急需提升^[1].针对当前手势识别方法在一些动作类似的手势识别中表现欠佳的问题,提出了一种多特征动态手势识别方法.该方法首先使用体感控制器Leap Motion追踪动态手势获取数据,然后在特征提取过程中增加对位移向量角度和拐点判定计数的提取,接着进行动态手势隐马尔科夫模型（Hidden Markov Model,HMM）的训练,最后根据待测手势与模型的匹配率进行识别.从实验结果中得出,该多特征识别方法能够提升相似手势的识别率.     

Using genetic programming for context-sensitive feature scoring in classification problems

Feature scoring is an avenue to feature selection that provides a measure of usefulness for the individual features of a classification task. Features are ranked based on their scores and selection is performed by choosing a small group of high-ranked features. Most existing feature scoring/ranking methods focus on the relevance of a single feature to the class labels regardless of the role of other features (context-insensitive). The paper proposes a genetic programming (GP)-based method to see how a set of features can contribute towards discriminating different classes. The features receive score in the context of other features participating in a GP program. The scoring mechanism is based on the frequency of appearance of each feature in a collection of GP programs and the fitness of those programs. Our results show that the proposed feature ranking method can detect important features of a problem. A variety of different classifiers restricted to just a few of these high-ranked features work well. The proposed scoring-ranking mechanism can also shrink the search space of size O(2 ⁿ ) of subsets of features to a search space of size O(n) in which there are points that are very likely to improve the classification performance.     

基于深度学习的采摘机器人目标识别定位算法

为了提升采摘机器人对果实的识别准确率以及定位定精度,提出一种基于深度学习Faster-RCNN框架的采摘机器人目标识别和定位算法.首先采用卷积神经网络VGG16模型提取输入图像的特性信息,并利用区域提议网络RPN生成含有目标的候选框,通过引入自适应候选框数的方法有效提升了算法性能,然后利用多任务损失函数对目标进行分类识...     

An Improved Method for Load Taxonomy Using Sample Shifting Technique and Signature Analysis

This paper illustrates an improved method of classification of electrical appliances, particularly for domestic loads, to construct load taxonomy on the basis of their signature analysis. Each electrical load is characterized by its own distinct signature and hence load signature analysis is useful in monitoring the health of the equipment, power quality, in determining individual energy usage etc. type of services. On the other hand, load taxonomy classifies these loads in several clusters on the basis of some features extracted from their signatures. In traditional methods of construction of load taxonomy, different signature patterns based on power metrics, V-I trajectories, Eigen vectors, etc. In this proposed method, with the adoption of sample shifting technique the required number of feature extraction is reduced to a lower value to find out various signature patterns than those are required in traditional load taxonomies. Moreover, a better taxonomy, having well separated groups of loads is achieved with lower number of extracted features.     

基于几何特征的飞机部位识别跟踪算法

针对复杂天气情况下形状可辨的飞机,提出了一种对飞机部位的识别跟踪算法。该算法利用飞机的几何特征,通过数学形态学的原理,在存有噪声干扰的运动图像中简单快速有效地识别出所需目标的重要部位,自动获取其跟踪模板,通过相关匹配技术实现对飞机重要部位的稳定跟踪。给出了应用该算法的具体步骤,并通过实验证明了该算法的有效性。     

基于听觉模型的客观音质评价方法研究

本文阐述了一种基于听觉模型的客观音质评价方法研究,我们在使用谱失真和谱包络失真进行了客观音质评价同时采用了听觉系统模拟的三个听觉感知模型,实验结果表明,在于听觉模型的谱夫真方法测量结果与主观评价更人一致性,在与主观评价具有同样相关系数时,采用听觉模型谱包失真方法的预测系数为４阶而ＬＰＣ倒谱距离方法的预测系数为８阶。     

用自产生和自组织神经网络对超声医学图像进行自动分割

本文研究用自产生和自组织神经网络方法进行超声心脏图像的自动分割。这种无监督的聚类方法能够自动搜索最佳的网络输出节点数而获取图像中的目标数,从而完成对图像的自动分割。实验结果表明,与自组织特征映射方法相比,本文的方法具有许多重要的优点。     

基于电容的手势识别系统

传统手势识别装置一般依赖于复杂计算设备,在简单场合配置手势识别装置成本过高。针对这一问题,提出了基于电容数字转换器的手势识别系统。具体实施过程中,电容数字转换器采集静态手势特征,滑动窗口滤波加工特征数值,再结合K邻近分类算法判别静态手势。该系统具有对硬件要求低、功耗小、对环境光强不敏感等优点。经过50次试验,系统对于常用的数字零到五手势识别正确率为98%,判别时间小于0.6 s。     

基于滚动引导滤波的红外与可见光图像融合算法

针对红外与可见光图像融合时,易产生细节丢失、噪声抑制不佳等问题,本文提出了一种改进的滚动引导滤波融合算法.该算法充分利用了滚动引导滤波边缘和局部亮度保持特性,在通过均值滤波将输入图像分解为基础层与细节层的基础上,结合滚动引导滤波与高斯滤波获取输入图像的显著图,利用不同尺度参数的引导滤波对显著图优化得到权重图,将权重图作为权重分别指导基础层与细节层的融合,最后联合融合后的子图重构得到融合图像.针对3类测试数据进行的融合实验表明,与非下采样轮廓波变换、基于引导滤波、基于显著性检测的两个尺度的图像融合等经典方法相比,本文方法得到的融合图像不但从主观视觉效果上细节信息更丰富、目标对比度加强,并且在非线性相关信息熵、相位一致性等6项客观评价指标上均具有较好的效果.     

特征结合和相关反馈技术在医学图像检索中的应用

本文在单一检索的基础上,对基于灰度直方图的颜色特征,基于小波变换的纹理特征和基于不变矩的形状特征进行融合.为了使用户能够参与检索过程,又引入了相关反馈机制,通过调整权值使得检索的结果最终满足用户的检索要求.最后分别给出基于单一特征,特征融合和相关反馈方法的查准率和查全率,并对试验结果进行分析.