Visual tracking based on Distribution Fields and online weighted multiple instance learning

This paper presents an improved multiple instance learning (MIL) tracker representing target with Distribution Fields (DFs) and building a weighted-geometric-mean MIL classifier. Firstly, we adopt DF layer as feature instead of traditional Haar-like one to model the target thanks to the DF specificity and the landscape smoothness. Secondly, we integrate sample importance into the weighted-geometric-mean MIL model and derive an online approach to maximize the bag likelihood by AnyBoost gradient framework to select the most discriminative layers. Due to the target model consisting of selected discriminative layers, our tracker is more robust while needing fewer features than the traditional Haar-like one and the original DFs one. The experimental results show higher performances of our tracker than those of five state-of-the-art ones on several challenging video sequences.     

多通道Haar-like特征多示例学习目标跟踪

目的 提出一种基于多通道Haar-like特征的多示例学习目标跟踪算法,克服了多示例跟踪算法在处理彩色视频时利用信息少和弱特征不能更换的缺点。方法 首先,针对原始多示例学习跟踪算法对彩色视频帧采用单通道信息或将其简单转化为灰度图像进行跟踪会丢失部分特征信息的缺点,提出在RGB三通道上生成位置、大小和通道完全随机的Haar-like特征来更好地表示目标。其次,针对多示例学习跟踪算法中Haar-like弱特征不能更换,难以反映目标自身和外界条件变化的特点,提出在弱分类器选择过程中,用随机生成的新Haar-like特征实时替换部分判别力最弱的Haar-like特征,从而在目标模型中引入新的信息,以适应目标外观的动态变化。结果 对8个具有挑战性的彩色视频序列的实验结果表明,与原始多示例学习跟踪算法、加权多示例学习跟踪算法、基于分布场的跟踪算法相比,提出的方法不仅获得了最小的平均中心误差,而且平均跟踪准确率比上述3种算法分别高52.85%,34.75%和5.71%,在4种算法中获得最优性能。结论 通过将Haar-like特征从RGB三通道随机生成,并将判别力最弱的部分Haar-like弱特征实时更换,显著提升了原始多示例学习跟踪算法对彩色视频的跟踪效果,扩展了其应用前景。     

Random set framework for multiple instance learning

Multiple instance learning (MIL) is a technique used for learning a target concept in the presence of noise or in a condition of uncertainty. While standard learning techniques present the learner with individual samples, MIL alternatively presents the learner with sets of samples. Although sets are the primary elements used for analysis in MIL, research in this area has focused on using standard analysis techniques. In the following, a random set framework for multiple instance learning (RSF-MIL) is proposed that can directly perform analysis on sets. The proposed method uses random sets and fuzzy measures to model the MIL problem, thus providing a more natural mathematical framework, a more general MIL solution, and a more versatile learning tool. Comparative experimental results using RSF-MIL are presented for benchmark data sets. RSF-MIL is further compared to the state-of-the-art in landmine detection using ground penetrating radar data.     

一种基于CRO的高阶神经网络多示例学习方法

在多示例学习(MIL)中,包是含有多个示例的集合,训练样本只给出包的标记,而没有给出单个示例的标记。提出一种基于示例标记强度的MIL方法(ILI-MIL),其允许示例标记强度为任何实数。考虑到基于梯度训练神经网络方法的计算复杂性和ILI-MIL目标函数的复杂性,利用基于化学反应优化的高阶神经网络来实现ILI-MIL,学习方法具有较强的非线性表达能力和较高的计算效率。实验结果表明,该算法比已有算法具有更加有效的分类能力,且适应范围更广。     

Effects of classifier structures and training regimes on integrated segmentation and recognition of handwritten numeral strings

In integrated segmentation and recognition of character strings, the underlying classifier is trained to be resistant to noncharacters. We evaluate the performance of state-of-the-art pattern classifiers of this kind. First, we build a baseline numeral string recognition system with simple but effective presegmentation. The classification scores of the candidate patterns generated by presegmentation are combined to evaluate the segmentation paths and the optimal path is found using the beam search strategy. Three neural classifiers, two discriminative density models, and two support vector classifiers are evaluated. Each classifier has some variations depending on the training strategy: maximum likelihood, discriminative learning both with and without noncharacter samples. The string recognition performances are evaluated on the numeral string images of the NIST special database 19 and the zipcode images of the CEDAR CDROM-1. The results show that noncharacter training is crucial for neural classifiers and support vector classifiers, whereas, for the discriminative density models, the regularization of parameters is important. The string recognition results compare favorably to the best ones reported in the literature though we totally ignored the geometric context. The best results were obtained using a support vector classifier, but the neural classifiers and discriminative density models show better trade-off between accuracy and computational overhead.     

基于区分性准则的Bottleneck特征及其在LVCSR中的应用

基于深层神经网络中间层的Bottleneck(BN)特征由于可以采用传统的混合高斯模型-隐马尔可夫建模(Gaussian mixture model-hidden Markov model, GMM-HMM),在大规 模连续语音识别中获得了广泛的应用。为了提取区分性的BN特征,本文提出在使用传统的BN特征训练好GMM-HMM模型之后,利用最小音素错误率（Minimum phone error, MPE）准则来优化BN网络参数以及GMM-HMM模型参数。该算法相对于其他区分性训练算法而言,采用的是全部数据作为一个大的数据包,而 不是小的包方式来训练深度神经网络,从而可以大大加快训练速度。实验结果表明,优化后的BN特征提取网络比传统方法能获得9%的相对词错误率下降。     

基于AFSVM-MIL算法的图像标注*

通常情况下关键字只标注在图像上,而多示例(MIL)检索的需要将关键字下沉到区域.针对这个问题,在模糊支持向量机算法(FSVM)的基础上提出了一种改进的自适应模糊支持向量机多示例学习算法(AFSVM-MIL算法),在多示例学习的框架下把区域级的图像标注变成了一种有监督的学习.该方法利用AFSVMMIL对训练集进行分类,结...     

Learning group-based dictionaries for discriminative image representation

Dictionary learning is a critical issue for achieving discriminative image representation in many computer vision tasks such as object detection and image classification. In this paper, a new algorithm is developed for learning discriminative group-based dictionaries, where the inter-concept (category) visual correlations are leveraged to enhance both the reconstruction quality and the discrimination power of the group-based discriminative dictionaries. A visual concept network is first constructed for determining the groups of visually similar object classes and image concepts automatically. For each group of such visually similar object classes and image concepts, a group-based dictionary is learned for achieving discriminative image representation. A structural learning approach is developed to take advantage of our group-based discriminative dictionaries for classifier training and image classification. The effectiveness and the discrimination power of our group-based discriminative dictionaries have been evaluated on multiple popular visual benchmarks.     

基于稀疏表达的多示例学习目标追踪算法

追踪目标在经历较大姿势变化时,会导致追踪目标偏移甚至丢失。为此,提出一种基于稀疏表达的多示例学习目标追踪算法。联合多示例学习与稀疏表达方法,将目标物体的局部稀疏编码作为多示例学习的训练数据,通过学习正负样本的局部稀疏编码获得一个多示例学习的分类器,分类的结果与粒子滤波框架相结合,估计目标在整个视频序列中的运动状态。实验结果表明,该算法稳定性较好,与增量学习追踪算法、范式学习追踪算法和多示例学习追踪算法相比,其中心位置误差率减少30%以上。     

Evaluation of Localized Semantics: Data, Methodology, and Experiments

We present a new data set of 1014 images with manual segmentations and semantic labels for each segment, together with a methodology for using this kind of data for recognition evaluation. The images and segmentations are from the UCB segmentation benchmark database (Martin et al., in International conference on computer vision, vol. II, pp. 416–421, 2001). The database is extended by manually labeling each segment with its most specific semantic concept in WordNet (Miller et al., in Int. J. Lexicogr. 3(4):235–244, 1990). The evaluation methodology establishes protocols for mapping algorithm specific localization (e.g., segmentations) to our data, handling synonyms, scoring matches at different levels of specificity, dealing with vocabularies with sense ambiguity (the usual case), and handling ground truth regions with multiple labels. Given these protocols, we develop two evaluation approaches. The first measures the range of semantics that an algorithm can recognize, and the second measures the frequency that an algorithm recognizes semantics correctly. The data, the image labeling tool, and programs implementing our evaluation strategy are all available on-line (kobus.ca//research/data/IJCV_2007). We apply this infrastructure to evaluate four algorithms which learn to label image regions from weakly labeled data. The algorithms tested include two variants of multiple instance learning (MIL), and two generative multi-modal mixture models. These experiments are on a significantly larger scale than previously reported, especially in the case of MIL methods. More specifically, we used training data sets up to 37,000 images and training vocabularies of up to 650 words. We found that one of the mixture models performed best on image annotation and the frequency correct measure, and that variants of MIL gave the best semantic range performance. We were able to substantively improve the performance of MIL methods on the other tasks (image annotation and frequency correct region labeling) by providing an appropriate prior.     

Multiple instance learning with bag dissimilarities

Multiple instance learning (MIL) is concerned with learning from sets (bags) of objects (instances), where the individual instance labels are ambiguous. In this setting, supervised learning cannot be applied directly. Often, specialized MIL methods learn by making additional assumptions about the relationship of the bag labels and instance labels. Such assumptions may fit a particular dataset, but do not generalize to the whole range of MIL problems. Other MIL methods shift the focus of assumptions from the labels to the overall (dis)similarity of bags, and therefore learn from bags directly. We propose to represent each bag by a vector of its dissimilarities to other bags in the training set, and treat these dissimilarities as a feature representation. We show several alternatives to define a dissimilarity between bags and discuss which definitions are more suitable for particular MIL problems. The experimental results show that the proposed approach is computationally inexpensive, yet very competitive with state-of-the-art algorithms on a wide range of MIL datasets.     

Incorporating multiple SVMs for automatic image annotation

In this paper, a novel automatic image annotation system is proposed, which integrates two sets of support vector machines (SVMs), namely the multiple instance learning (MIL)-based and global-feature-based SVMs, for annotation. The MIL-based bag features are obtained by applying MIL on the image blocks, where the enhanced diversity density (DD) algorithm and a faster searching algorithm are applied to improve the efficiency and accuracy. They are further input to a set of SVMs for finding the optimum hyperplanes to annotate training images. Similarly, global color and texture features, including color histogram and modified edge histogram, are fed into another set of SVMs for categorizing training images. Consequently, two sets of image features are constructed for each test image and are, respectively, sent to the two sets of SVMs, whose outputs are incorporated by an automatic weight estimation method to obtain the final annotation results. Our proposed annotation approach demonstrates a promising performance for an image database of 12 000 general-purpose images from COREL, as compared with some current peer systems in the literature.     

A discriminative structural model for joint segmentation and recognition of human actions

Achieving joint segmentation and recognition of continuous actions in a long-term video is a challenging task due to the varying durations of actions and the complex transitions of multiple actions. In this paper, a novel discriminative structural model is proposed for splitting a long-term video into segments and annotating the action label of each segment. A set of state variables is introduced into the model to explore discriminative semantic concepts shared among different actions. To exploit the statistical dependences among segments, temporal context is captured at both the action level and the semantic concept level. The state variables are treated as latent information in the discriminative structural model and inferred during both training and testing. Experiments on multi-view IXMAS and realistic Hollywood datasets demonstrate the effectiveness of the proposed method.     

MILD: Multiple-Instance Learning via Disambiguation

In multiple-instance learning (MIL), an individual example is called an instance and a bag contains a single or multiple instances. The class labels available in the training set are associated with bags rather than instances. A bag is labeled positive if at least one of its instances is positive; otherwise, the bag is labeled negative. Since a positive bag may contain some negative instances in addition to one or more positive instances, the true labels for the instances in a positive bag may or may not be the same as the corresponding bag label and, consequently, the instance labels are inherently ambiguous. In this paper, we propose a very efficient and robust MIL method, called Multiple-Instance Learning via Disambiguation (MILD), for general MIL problems. First, we propose a novel disambiguation method to identify the true positive instances in the positive bags. Second, we propose two feature representation schemes, one for instance-level classification and the other for bag-level classification, to convert the MIL problem into a standard single-instance learning (SIL) problem that can be solved by well-known SIL algorithms, such as support vector machine. Third, an inductive semi-supervised learning method is proposed for MIL. We evaluate our methods extensively on several challenging MIL applications to demonstrate their promising efficiency, robustness, and accuracy.     

面向层次分类的文本特征选择方法

提出一种针对层次分类的文本特征选择方法。先给出类别层次相关度的概念,并利用分类树和训练数据在不同层次上的概率分布进行计算,进而得到分类树中不同类别的重要性。最后基于前面的计算结果,计算每个特征对类别的识别能力,并选择识别能力大的特征组成用于分类的特征集合。实验表明该方法在选取的特征质量以及在accuracy、F1和micro-Precision等分类测度上均优于传统方法。     

Object recognition using proportion-based prior information: Application to fisheries acoustics

This paper addresses the inference of probabilistic classification models using weakly supervised learning. The main contribution of this work is the development of learning methods for training datasets consisting of groups of objects with known relative class priors. This can be regarded as a generalization of the situation addressed by Bishop and Ulusoy (2005), where training information is given as the presence or absence of object classes in each set. Generative and discriminative classification methods are conceived and compared for weakly supervised learning, as well as a non-linear version of the probabilistic discriminative models. The considered models are evaluated on standard datasets and an application to fisheries acoustics is reported. The proposed proportion-based training is demonstrated to outperform model learning based on presence/absence information and the potential of the non-linear discriminative model is shown.     

Integrating models of discrimination and characterization

It is argued that in applications of concept learning from examples where not every possible category of the domain is present in the training set (i.e., many real world applications), classification performance can be improved by integrating suitable discriminative and characteristic models of classification. The suggested approach is to first discriminate between the categories present in the training set and then characterize each of these categories against all possible categories. To show the viability of this approach, a number of different discriminators and characterizers are integrated and tested. In particular, a novel characterization method that makes use of the information about the statistical distribution of feature values that can be extracted from the training examples is used. By using this method it is possible to control the degree of generalization and to deal with dependencies among features.     

基于改进初始化判别K SVD方法的人脸识别

基于稀疏表示的人脸识别问题希望字典同时具有良好的表示能力和较强的辨识性。采用判别式K SVD（D ksvd）算法,可训练得到较好的字典和线性判别函数,但该算法中的初始化字典是从各类样本中选择部分样本经K SVD方法得到的,不能较完整地表示所有样本的特性,影响了基于该初始字典的训练字典的表示能力和分类器的辨识性。在字典初始化方法上进行了改进,先训练类内字典再级联成新的初始化字典,由于类内训练字典是各类别的优化字典,降低了训练字典的误差,提高了训练字典与线性分类器的判别性,在保持较快识别速度的同时,提高了人脸识别率。     

Discriminative Training of the Hidden Vector State Model for Semantic Parsing

In this paper, we discuss how discriminative training can be applied to the Hidden Vector State (HVS) model in different task domains. The HVS model is a discrete Hidden Markov Model (HMM) in which each HMM state represents the state of a push-down automaton with a finite stack size. In previous applications, Maximum Likelihood estimation (MLE) is used to derive the parameters of the HVS model. However, MLE makes a number of assumptions and unfortunately some of these assumptions do not hold. Discriminative training, without making such assumptions, can improve the performance of the HVS model. Experiments have been conducted in two domains: the travel domain for the semantic parsing task using the DARPA Communicator data and the ATIS data, and the bioinformatics domain for the information extraction task using the GENIA corpus. The results demonstrate modest improvements of the performance of the HVS model using discriminative training. In the travel domain, discriminative training of the HVS model gives a relative error reduction rate of 31% in F-measure when compared with MLE on the DARPA Communicator data and 9% on the ATIS data. In the bioinformatics domain, a relative error reduction rate of 4% in F-measure is achieved on the GENIA corpus.