基于F-score特征选择和支持向量机的P300识别算法

如何从脑电信号中快速准确地识别出P300成分是脑-机接口研究中的一个热点问题.针对P300的识别问题,我们提出了一种将F-score特征选择与支持向量机相结合的判别方法,该方法采用F-score特征选择减少输入特征的维数,以克服支持向量机算法判别速度慢的缺点;然后借助支持向量机算法良好的分类性能实现P300的识别.本文在BCI Competition 2003的P300实验数据集上对该方法进行了验证,结果表明,在5次重复实验中该方法的识别准确率达到了100%,且判别速度与未经特征选择的传统支持向量机算法相比提高了近2倍.     

一种运动想象脑电分类算法的研究

为了解决脑机接口(BCI)中不同意识任务下脑电信号分类问题,针对运动想象脑电(EEG)的事件相关去同步/同步(ERD/ERS)现象,提出一种基于支持向量机(SVM)的实用分类算法。该算法首先对脑电信号进行滤波,获得对运动想象比较敏感的频段,对滤波后的脑电信号,通过去均值减小由于均值不同所造成的误差,然后,再提取基于ERD/ERS的脑电能量场强特征,对提取的特征,运用支持向量机(SVM)进行分类,得到了满意的效果。结果表明,此方法可为脑机接口技术的应用提供有效的手段。     

Emotion recognition from physiological signals

Emotion recognition is one of the great challenges in human–human and human–computer interaction. Accurate emotion recognition would allow computers to recognize human emotions and therefore react accordingly. In this paper, an approach for emotion recognition based on physiological signals is proposed. Six basic emotions: joy, sadness, fear, disgust, neutrality and amusement are analysed using physiological signals. These emotions are induced through the presentation of International Affecting Picture System (IAPS) pictures to the subjects. The physiological signals of interest in this analysis are: electromyogram signal (EMG), respiratory volume (RV), skin temperature (SKT), skin conductance (SKC), blood volume pulse (BVP) and heart rate (HR). These are selected to extract characteristic parameters, which will be used for classifying the emotions. The SVM (support vector machine) technique is used for classifying these parameters. The experimental results show that the proposed methodology provides in general a recognition rate of 85% for different emotional states.     

Breast Ultrasound Image Classification Based on Multiple-Instance Learning

Breast ultrasound (BUS) image segmentation is a very difficult task due to poor image quality and speckle noise. In this paper, local features extracted from roughly segmented regions of interest (ROIs) are used to describe breast tumors. The roughly segmented ROI is viewed as a bag. And subregions of the ROI are considered as the instances of the bag. Multiple-instance learning (MIL) method is more suitable for classifying breast tumors using BUS images. However, due to the complexity of BUS images, traditional MIL method is not applicable. In this paper, a novel MIL method is proposed for solving such task. First, a self-organizing map is used to map the instance space to the concept space. Then, we use the distribution of the instances of each bag in the concept space to construct the bag feature vector. Finally, a support vector machine is employed for classifying the tumors. The experimental results show that the proposed method can achieve better performance: the accuracy is 0.9107 and the area under receiver operator characteristic curve is 0.96 (p < 0.005).     

一种基于SVM的胶囊内窥镜出血智能识别方法

由于Wireless Capsule Endoscopy(WCE)在消化道中采集到的巨大数量的图像均需要医务人员靠肉眼来排查,给医生带来巨大的负担。该文提供了一种基于支持向量机(Support Vector Machine,SVM)分类器的胶囊内窥镜出血智能识别方法,创立一种新的特征参数,并对SVM参数的选择进行实验优化,最终达到94%的特异度与83%的灵敏度。     

改进的快速FCM及SVM实现糖网白色病灶的自动检测

为构建基于眼底图像的糖尿病视网膜病变(糖网)自动筛查系统,提出一种基于改进的快速FCM(IFFCM)及SVM的糖网白色病灶自动检测算法.首先,利用改进的快速FCM算法,对彩色眼底图像进行粗分割获取糖网白色病灶候选区域,由于该算法将中值滤波添加到FCM算法的准则函数中,同时利用K-means算法的聚类结果对FCM进行聚类中心初始化,使得该算法克服了传统FCM算法计算复杂度高以及对噪声敏感的缺点;其次,采用两层级联分类结构的SVM对候选区域进行分类,即先利用SVM根据候选区域的特征集将白色病灶提取出来,再利用SVM根据另外的特征集将白色病灶中的硬性渗出与棉绒斑区分开,从而实现眼底图像中糖网白色病灶的自动检测.利用该方法对65幅眼底图像进行糖网白色病灶的自动检测,得到图像水平灵敏度100％,特异性95.0％,准确率98.46％;病灶区域水平(硬性渗出/棉绒斑)灵敏度96.42％/97.15％,阳性预测值90.03％/91.18％;平均一幅图像处理时间35.56 s.结果表明:将改进的快速FCM算法所提供的良好粗分割结果与识别率较高的分类器SVM相结合,使得对糖网白色病灶的自动检测结果较优,即该算法能够高效地自动检测出眼底图像中的糖网白色病灶.     

Multi-level basis selection of wavelet packet decomposition tree for heart sound classification

Wavelet packet transform decomposes a signal into a set of orthonormal bases (nodes) and provides opportunities to select an appropriate set of these bases for feature extraction. In this paper, multi-level basis selection (MLBS) is proposed to preserve the most informative bases of a wavelet packet decomposition tree through removing less informative bases by applying three exclusion criteria: frequency range, noise frequency, and energy threshold. MLBS achieved an accuracy of 97.56% for classifying normal heart sound, aortic stenosis, mitral regurgitation, and aortic regurgitation. MLBS is a promising basis selection to be suggested for signals with a small range of frequencies.     

Determination of Impact Damage in CFRP via PVDF Signal Analysis with Support Vector Machine

Carbon fiber reinforced plastics (CFRPs) have high specific stiffness and strength, but they are vulnerable to transverse loading, especially low-velocity impact loadings. The impact damage may cause serious strength reduction in CFRP structure, but the damage in a CFRP is mainly internal and microscopic, that it is barely visible. Therefore, this study proposes a method of determining impact damage in CFRP via poly(vinylidene fluoride) (PVDF) sensor, which is convenient and has high mechanical and electrical performance. In total, 114 drop impact tests were performed to investigate on impact responses and PVDF signals due to impacts. The test results were analyzed to determine the damage of specimens and signal features, which are relevant to failure mechanisms were extracted from PVDF signals by means of discrete wavelet transform (DWT). Support vector machine (SVM) was used for optimal classification of damage state, and the model using radial basis function (RBF) kernel showed the best performance. The model was validated through a 4-fold cross-validation, and the accuracy was reported to be 92.30%. In conclusion, impact damage in CFRP structures can be effectively determined using the spectral analysis and the machine learning-based classification on PVDF signals.