基于支持向量机增量学习和LPBoost的人体目标再识别算法

摄像机间目标关联是无重叠视域多摄像机目标持续跟踪的关键.提出了一种只利用人体目标外观,完全不依赖于空时关系的人体目标再识别算法,利用识别结果直接进行跨摄像机间人体目标关联,而不依赖于目标的捕获时间和路径限制.对跟踪视频前景图像序列提取互补性视觉单词树直方图和全局颜色直方图二种特征,采用支持向量机增量学习在线训练二种特征的人体外观辨别模型,再利用多类线性规划增强算法对二种特征的支持向量机模型进行在线自适应融合.实验结果表明,本文算法具有较强的在线学习能力,能增量式表达人体目标辨别性外观模型,特征融合后的模型区别性更强,有效地降低多方面条件变化的影响,获得了高识别率,且能够实现快速实时实现,相对于现有方法有了明显提升.     

基于在线检测和尺度自适应的相关滤波跟踪

针对相关滤波跟踪在遮挡及目标尺度变化等情况下容易跟踪失败的问题,提出一种基于在线检测和尺度自适应的相关滤波跟踪算法。相关滤波跟踪器融合方向梯度直方图特征、颜色属性特征和光照不变特征进行目标定位;通过局部稀疏表示模型的重构残差进行遮挡判别,如果发生遮挡则进行在线支持向量机检测,实现目标重定位;进行由粗至精的尺度估计,通过尺度预估计和牛顿迭代法得到目标的精确尺度。采用均衡的模型更新策略,固定更新相关滤波器,保守更新稀疏表示模型和支持向量机。实验结果表明:与现有跟踪算法相比,所提算法能有效降低遮挡、目标尺度变化等复杂因素的干扰,并在50组测试序列上取得较高的距离精度和成功率,其整体性能优于其他对比算法。     

基于动态目标建模的粒子滤波视觉跟踪算法

提出一种根据场景变化动态建立目标模型的粒子滤波视觉跟踪算法.该方法首先选择简单且具有互补性的特征描述当前图像,并统一采用直方图法对这些特征进行建模;然后在粒子滤波框架下,根据巴塔恰里亚测度评价各个目标特征和背景特征之间的可区分程度,动态调整特征间的置信度;并对各个特征似然函数的噪音参量进行在线估计和更新,使其似然函数的度量标准达到统一.分析和实验表明,该算法性能优于仅仅采用多特征融合进行粒子滤波视觉跟踪的方法,对摄像机运动、混淆干扰、遮挡及目标外观大小的改变具有更强的鲁棒性.     

基于动态特征融合的粒子滤波目标跟踪算法

提出一种基于动态特征融合的粒子滤波目标跟踪算法。选择具有互补性的灰度直方图和梯度直方图特征共同描述目标模型,然后在目标跟踪过程中,根据特征对目标和背景的区分程度动态地调整每个特征的置信度,对目标模型进行在线动态建模和更新,从而提高目标模型描述的准确度,并进一步提高粒子滤波算法的跟踪精度。实验结果表明:在对典型场景下的目标跟踪过程中,提出的算法比单独使用一种特征的粒子滤波算法具有更高的跟踪精度和更稳定可靠的跟踪性能。     

基于特征融合的长时目标跟踪算法

针对长时目标跟踪中目标遮挡、目标出视野等因素导致的目标失跟问题,提出一种基于特征融合的长时目标跟踪算法,提高目标跟踪的速度和稳健性。首先,融合目标方向梯度直方图特征、颜色空间特征和局部敏感直方图特征,来增强算法在复杂情况下的特征判别力,提高目标跟踪的稳健性,并对融合特征进行降维来提高目标跟踪的速度;然后,通过额外的一维尺度相关滤波器来获得目标最优的尺度估计,并通过正交三角分解来无损降低计算复杂度;最后,自适应确定目标检测阈值,在目标遮挡或出视野导致目标失跟时,通过EdgeBoxes方法提取目标候选区域,利用结构化支持向量机重新检测目标位置达到长时跟踪的目的。在标准跟踪数据集OTB2015和UAV123上进行实验。结果表明,本文算法较对比算法中最优算法目标跟踪平均精度提升5.0%,目标跟踪平均成功率提升2.6%,目标跟踪平均速度为28.2 frame/s,可满足跟踪的实时性要求。在目标受到遮挡、出视野等情况下,该算法仍能够对目标进行持续准确的跟踪。     

基于多级跟踪队列的运动目标跟踪遮挡处理

针对多目标跟踪过程中存在的遮挡问题,提出了一种固定摄像机场景下的多目标实时跟踪算法.提出基于鬼影判别与背景模型选择更新的背景差法检测运动目标,建立一种融合色度与边缘特征的目标模型.通过定义稳定跟踪队列、临时跟踪队列、跟踪丢失队列以及候选跟踪队列等跟踪器队列,提出基于多级关联匹配的策略实现多目标跟踪遮挡处理,针对新目标、...     

基于颜色和空间信息的多特征融合目标跟踪算法

为解决单一特征目标跟踪鲁棒性较差的问题,提出一种基于颜色和空间信息的多特征融合目标跟踪算法。采用一种自适应划分颜色区间的方法提取目标颜色特征,利用空间直方图提取目标颜色的空间分布信息。在粒子滤波框架下将自适应颜色直方图和空间直方图相结合,在特征融合中引入特征不确定性度量方法,自适应调整不同特征对跟踪结果的贡献,提高算法的鲁棒性。仿真实验结果表明,该跟踪算法平均位置最小误差值仅6.967 像素,而单一特征跟踪算法以及传统融合算法的跟踪误差达192.576 像素和199.464像素。说明本文算法在跟踪准确性上优于单一特征跟踪算法及传统融合算法,具有更好的跟踪精度和更高的鲁棒性。     

融合声纹信息的能量谱图在鸟类识别中的研究

常用的梅尔倒谱系数结合高斯混合模型(MFCC+GMM)方法的鸟鸣声识别技术难适应噪声环境,模型难以收敛,且计算复杂度高。该文提出一种融合声纹信息的能量谱图的鸟类识别方法(VPS-BR),该方法利用鸟类鸣声在能量谱图上所表现的多维差异性,定量识别鸣声声纹特征。通过对分贝能量进行颜色映射得到能量谱图,提取其视觉特征所表达的声学特征,分析归纳得到鸟类特有鸣声模式。在特征提取步骤中,选用识别速度快的局部二值模式、识别鲁棒性高的方向梯度直方图两个参数表征鸟鸣声谱图的边缘声纹;在识别步骤中,用局部二值模式和方向梯度直方图两种特征分别与支持向量机、K最近邻和随机森林3种分类器算法进行两两组合构建识别模型测试。对15种原始带噪鸟类鸣声数据集进行交叉验证,VPS-BR模型的平均识别率比MFCC+GMM组合模型高出11.3%,方向梯度直方图特征与K最近邻分类器的组合模型识别率达90.5%,表现出较好的抗噪性能和识别性能。最后针对样本数据集缺乏问题,使用生成对抗网络进行图像增强,进一步将识别率提升1.48%。     

复杂红外地面环境下的稳定目标跟踪方法

针对红外目标相关滤波跟踪过程中由于背景杂波干扰、目标遮挡和目标形变等情况导致的鲁棒性差甚至跟踪目标丢失的问题,提出一种融合跟踪-学习-检测方法和相关滤波理论的红外目标跟踪算法.该算法在传统相关滤波框架基础上,融合目标的方向梯度直方图特征和亮度直方图特征,改善了目标轻微形变导致的模型漂移问题.针对背景杂波和遮挡导致的多峰值响应问题,对目标背景区域的相关响应进行惩罚,建立目标和背景响应的多模态检测机制,实现目标由粗到精的定位,并采用自适应的学习率优化跟踪模型的漂移问题;针对目标被严重遮挡或脱离视野的问题,通过全局目标再检测,实现目标的重捕.实验结果表明,在复杂红外地面环境下,该算法有效地解决了相似目标干扰和目标被严重遮挡导致的目标丢失问题.基于OTB-2015视频基准序列和红外视频序列测试,对比多个主流的相关滤波跟踪算法,该算法在跟踪精度和成功率方面较长时相关滤波跟踪算法分别提升了5.6%和4.1%;在目标遮挡指标测试中,该算法在跟踪精度和成功率方面相较长时相关滤波跟踪算法分别提升了4.6%和6.1%.     

基于支持向量机的舰船图像识别

支持向量机(SVM)是一种基于超平面分类的新的学习方法,具有很强的泛化能力。研究了支持向量机的学习机理,以及实现支持向量机的序贯最小优化算法(SMO),并用来对舰船图像进行识别。首先将待识别目标进行二维小波分解,获取不同尺度下的小波系数,然后对其进行主元分析,得到的主元分量作为支持向量机的特征量输入。实验结果表明,该方法具有良好的分类性能。     

A semi-supervised learning approach for detection of phishing webpages

This paper proposes a new phishing webpage detection approach based on a kind of semi-supervised learning method-transductive support vector machine (TSVM). Firstly the features of web image are extracted for complementing the disadvantage of phishing detection only based on document object model (DOM); they include gray histogram, color histogram, and spatial relationship between subgraphs. Then the features of sensitive information are examined by using page analysis based on DOM objects. In contrast to the drawback of support vector machine (SVM) algorithm which simply trains classifier by learning little and poor representative labeled samples, this method introduces the TSVM to train classifier that it takes into account the distribution information implicitly embodied in the large quantity of the unlabeled samples, and have better performance than SVM. The experimental results show that the proposed method not only achieves better classification accuracy, but also has strong applicability as the independent method of phishing detection.     

层叠式“产生/判别”混合模型的语音情感识别

提出了层叠式“产生/判别”混合模型的语音情感识别方法。首先,提取63维语句级特征,运用Fisher从中选择12个最佳的语句级特征,建立小波神经网络(WNN)的层叠式产生式模型进行语音情感识别;然后提取69维帧级特征,采用SFS选择出待使用的8维特征,将高斯混合模型(GMM)进行多维概率输出,建立层叠式“产生/判别”混合模型进行语音情感识别。实验结果显示:(1)层叠式“产生/判别”混合模型较单独WNN、GMM、HMM (隐马尔可夫模型)、SVM (支持向量机)的识别率要高;(2)层叠式“产生/判决式”混合模型识别率较基于WNN的层叠产生式模型高;(3) M=13,D维GMM-MAP/SVM (MAP,最大后验概率)串联融合模型为最优的层叠式“产生/判别”混合模型,能获得最高85.1%的识别率。      

Accurate object detection with a discriminative shape model

Discriminative model over bag-of-visual-words representation significantly improves the accuracy of object detection under clutter. However, it encounters bottleneck because of completely ignoring geometric constraint between features. On the contrary, to detect object accurately explicit shape model heavily relies on geometric information of the object, which as a result lacks of discriminative power. In this paper, we present a discriminative shape model to make use of the advantages of the two models based on the insight that the two models are essentially complementary. Discriminative model provides discriminative power, while shape model encodes geometry. The cost function that we used to distinguish objects considers both the detection maps of the discriminative model and the result of shape matching. In this cost function, we adopt a novel way to deal with multi-scale detection maps. We show that this cost function has very strong discriminative power, which makes learning a discriminative threshold for full object detection possible. For shape model, we also present a scheme for learning a good shape model from noisy images. Experiments on UIUC Car and Weizmann–Shotton horses show state-of-the-art performance of our model.     

Automatic knee cartilage segmentation from multi-contrast MR images using support vector machine classification with spatial dependencies

Accurate segmentation of knee cartilage is required to obtain quantitative cartilage measurements, which is crucial for the assessment of knee pathology caused by musculoskeletal diseases or sudden injuries. This paper presents an automatic knee cartilage segmentation technique which exploits a rich set of image features from multi-contrast magnetic resonance (MR) images and the spatial dependencies between neighbouring voxels. The image features and the spatial dependencies are modelled into a support vector machine (SVM)-based association potential and a discriminative random field (DRF)-based interaction potential. Subsequently, both potentials are incorporated into an inference graphical model such that the knee cartilage segmentation is cast into an optimal labelling problem which can be efficiently solved by loopy belief propagation. The effectiveness of the proposed technique is validated on a database of multi-contrast MR images. The experimental results show that using diverse forms of image and anatomical structure information as the features are helpful in improving the segmentation, and the joint SVM-DRF model is superior to the classification models based solely on DRF or SVM in terms of accuracy when the same features are used. The developed segmentation technique achieves good performance compared with gold standard segmentations and obtained higher average DSC values than the state-of-the-art automatic cartilage segmentation studies.     

Tone model integration based on discriminative weight training for Putonghua speech recognition

A discriminative framework of tone model integration in continuous speech recognition was proposed. The method uses model dependent weights to scale probabilities of the hidden Markov models based on spectral features and tone models based on tonal features. The weights are discriminatively trained by minimum phone error criterion. Update equation of the model weights based on extended Baum-Welch algorithm is derived. Various schemes of model weight combination are evaluated and a smoothing technique is introduced to make training robust to over fitting. The proposed method is ewluated on tonal syllable output and character output speech recognition tasks. The experimental results show the proposed method has obtained 9.5% and 4.7% relative error reduction than global weight on the two tasks due to a better interpolation of the given models. This proves the effectiveness of discriminative trained model weights for tone model integration.     

Three-class classification models of logS and logP derived by using GA–CG–SVM approach

In this investigation, three-class classification models of aqueous solubility (logS) and lipophilicity (logP) have been developed by using a support vector machine (SVM) method combined with a genetic algorithm (GA) for feature selection and a conjugate gradient method (CG) for parameter optimization. A 5-fold cross-validation and an independent test set method were used to evaluate the SVM classification models. For logS, the overall prediction accuracy is 87.1% for training set and 90.0% for test set. For logP, the overall prediction accuracy is 81.0% for training set and 82.0% for test set. In general, for both logS and logP, the prediction accuracies of three-class models are slightly lower by several percent than those of two-class models. A comparison between the performance of GA–CG–SVM models and that of GA–SVM models shows that the SVM parameter optimization has a significant impact on the quality of SVM classification model. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users. Hui Zhang and Ming-Li Xiang are contributed equally.     

高光谱无损识别野生和种植黑枸杞

高光谱图像技术在农产品检测及识别方面有广阔的应用前景。野生黑枸杞经济效益显著,经常被种植黑枸杞冒充。提出一种利用高光谱图像对野生黑枸杞无损快速识别的方法。主要内容和结果如下：（1）共采集256份（野生、种植各128份）黑枸杞在900～1 700 nm范围的高光谱反射光谱,每份平均光谱作为此样品的光谱;（2）采用标准正态变换（SNV）对采集的光谱预处理;基于Kennard-Stone法,按照校正集和预测集比例为2∶1对样品划分,用连续投影算法（SPA）对光谱进行降维处理,提取特征波长30个;分别将全光谱和SPA 提取的30个特征波长作为模型输入,建立支持向量机（SVM）、极限学习机（ELM）和随机森林（RF）识别模型。（3）结果表明,在识别野生黑枸杞模型中,基于全光谱和SPA建立的SVM,ELM和RF模型校正集识别率均高于98.8%,基于全光谱和SPA建立的SVM,ELM和RF模型预测集识别率均高于97.7%。基于全光谱（FS）建立的三种识别模型略优于基于SPA建立的三种识别模型。但从简化模型方面,SPA提取的特征波常数仅为全光谱的11.8%,大大降低了模型运算量。三种模型中,基于随机森林模型无损识别野生黑枸杞效果最好,均达到100%。研究表明,利用高光谱图像技术结合分类模型可快速识别野生黑枸杞。     

An object tracking method based on guided filter for night fusion image

Online object tracking is a challenging problem as it entails learning an effective model to account for appearance change caused by intrinsic and extrinsic factors. In this paper, we propose a novel online object tracking with guided image filter for accurate and robust night fusion image tracking. Firstly, frame difference is applied to produce the coarse target, which helps to generate observation models. Under the restriction of these models and local source image, guided filter generates sufficient and accurate foreground target. Then accurate boundaries of the target can be extracted from detection results. Finally timely updating for observation models help to avoid tracking shift. Both qualitative and quantitative evaluations on challenging image sequences demonstrate that the proposed tracking algorithm performs favorably against several state-of-art methods.