Probabilistic visual learning for object representation

We present an unsupervised technique for visual learning, which is based on density estimation in high-dimensional spaces using an eigenspace decomposition. Two types of density estimates are derived for modeling the training data: a multivariate Gaussian (for unimodal distributions) and a mixture-of-Gaussians model (for multimodal distributions). Those probability densities are then used to formulate a maximum-likelihood estimation framework for visual search and target detection for automatic object recognition and coding. Our learning technique is applied to the probabilistic visual modeling, detection, recognition, and coding of human faces and nonrigid objects, such as hands     

基于深度学习的视觉单目标跟踪综述

单目标跟踪是一种在视频中利用目标外观和上下文信息对单个目标分析运动状态、提供定位的技术,在智能监控、智能交互、导航制导等方面具有应用前景,但遮挡、背景干扰、目标变化等问题导致实际应用的进展缓慢.随着近年来深度学习的快速发展,研究使用深度学习技术优化单目标跟踪算法已成为计算机视觉领域的热点之一.围绕基于深度学习的单目标跟踪算法,在分析了单目标跟踪的基本原理基础上,从相关滤波、孪生网络、元学习、注意力、循环神经网络和生成对抗网络六个方面,根据核心算法的不同分别进行了概述和分析;此外,对研究现状进行了总结,提出了算法的发展趋势和优化思路.     

在线复合模板模型表示的视觉目标跟踪

目的 视觉目标跟踪中,目标往往受到自身或场景中各种复杂干扰因素的影响,这对正确捕捉所感兴趣的目标信息带来极大的挑战。特别是,跟踪器所用的模板数据主要是在线学习获得,数据的可靠性直接影响到候选样本外观模型表示的精度。针对视觉目标跟踪中目标模板学习和候选样本外观模型表示等问题,采用一种较为有效的模板组织策略以及更为精确的模型表示技术,提出一种新颖的视觉目标跟踪算法。方法 跟踪框架中,将候选样本外观模型表示假设为由一组复合模板和最小重构误差组成的线性回归问题,首先利用经典的增量主成分分析法从在线高维数据中学习出一组低维子空间基向量(模板正样本),并根据前一时刻跟踪结果在线实时采样一些特殊的负样本加以扩充目标模板数据,再利用新组织的模板基向量和独立同分布的高斯—拉普拉斯混合噪声来线性拟合候选目标外观模型,最后估计出候选样本和真实目标之间的最大似然度,从而使跟踪器能够准确捕捉每一时刻的真实目标状态信息。结果 在一些公认测试视频序列上的实验结果表明,本文算法在目标模板学习和候选样本外观模型表示等方面比同类方法更能准确有效地反映出视频场景中目标状态的各种复杂变化,能够较好地解决各种不确定干扰因素下的模型退化和跟踪漂移问题,和一些优秀的同类算法相比,可以达到相同甚至更高的跟踪精度。结论 本文算法能够在线学习较为精准的目标模板并定期更新,使得跟踪器良好地适应内在或外在因素(姿态、光照、遮挡、尺度、背景扰乱及运动模糊等)所引起的视觉信息变化,始终保持其最佳的状态,使得候选样本外观模型的表示更加可靠准确,从而展现出更为鲁棒的性能。     

基于多任务混合噪声分布模型表示的视频跟踪

视觉跟踪中,目标信息是不确定的非线性变化过程。随时间和空间而变化的复杂动态数据中学习出较为精确的目标模板并用它来线性表示候选样本外观模型,从而使跟踪器较好地适应跟踪作业中内在或外在因素所引起的目标外观变化是视觉目标跟踪研究的重点。提出一种新颖的多任务混合噪声分布模型表示的视频跟踪算法,将候选样本外观模型假设为由一组目标模板和最小重构误差组成的多任务线性回归问题。利用经典的增量主成分分析法从高维数据中学习出一组低维子空间基向量（模板正样本）,并在线实时采样一些特殊的负样本加以扩充目标模板,再利用扩充后的新模板和独立同分布的高斯-拉普拉斯混合噪声来线性拟合当前时刻的候选目标外观模型,最后计算候选样本和真实目标之间的最大似然度,从而准确捕捉当前时刻的真实目标。在一些公认测试视频上的实验结果表明,该算法将能够在线学习较为精准的目标模板并定期更新目标在不同状态时的特殊信息,使得跟踪器始终保持最佳的状态,从而良好地适应不断发生变化的视觉信息（姿态、光照、遮挡、尺度、背景扰乱及运动模糊等）,表现出更好的鲁棒性能。     

判别稀疏表示与在线字典学习的运动目标跟踪

针对传统稀疏表示不能有效区分目标和背景的缺点,提出一种判别稀疏表示算法,这种算法在传统稀疏表示目标函数中加入一个判别函数,大大降低干扰因素对目标跟踪的影响。基于判别稀疏表示和[?1]约束,提出一种在线字典学习算法升级目标模板,有效降低背景信息对目标模板的影响。提取目标梯度方向的直方图（HOG）特征,利用其对光照和形变等复杂环境具有较强鲁棒性的优点,实现对目标更稳定的跟踪。实验结果表明,与现有跟踪方法相比,该算法的跟踪效果更好。     

Collaborative strategy for visual object tracking

Multimedia Tools and Applications - Adaptively learning the difference between object and background, discriminative trackers are able to overcome the complex background problem in visual object...     

Online multi-view subspace learning via group structure analysis for visual object tracking

In this paper, we focus on incrementally learning a robust multi-view subspace representation for visual object tracking. During the tracking process, due to the dynamic background variation and target appearance changing, it is challenging to learn an informative feature representation of tracking object, distinguished from the dynamic background. To this end, we propose a novel online multi-view subspace learning algorithm (OMEL) via group structure analysis, which consistently learns a low-dimensional representation shared across views with time changing. In particular, both group sparsity and group interval constraints are incorporated to preserve the group structure in the low-dimensional subspace, and our subspace learning model will be incrementally updated to prevent repetitive computation of previous data. We extensively evaluate our proposed OMEL on multiple benchmark video tracking sequences, by comparing with six related tracking algorithms. Experimental results show that OMEL is robust and effective to learn dynamic subspace representation for online object tracking problems. Moreover, several evaluation tests are additionally conducted to validate the efficacy of group structure assumption.     

深度学习的目标跟踪算法综述

目标跟踪是利用一个视频或图像序列的上下文信息,对目标的外观和运动信息进行建模,从而对目标运动状态进行预测并标定目标位置的一种技术,是计算机视觉的一个重要基础问题,具有重要的理论研究意义和应用价值,在智能视频监控系统、智能人机交互、智能交通和视觉导航系统等方面具有广泛应用。大数据时代的到来及深度学习方法的出现,为目标跟踪的研究提供了新的契机。本文首先阐述了目标跟踪的基本研究框架,从观测模型的角度对现有目标跟踪的历史进行回顾,指出深度学习为获得更为鲁棒的观测模型提供了可能;进而从深度判别模型、深度生成式模型等方面介绍了适用于目标跟踪的深度学习方法;从网络结构、功能划分和网络训练等几个角度对目前的深度目标跟踪方法进行分类并深入地阐述和分析了当前的深度目标跟踪方法;然后,补充介绍了其他一些深度目标跟踪方法,包括基于分类与回归融合的深度目标跟踪方法、基于强化学习的深度目标跟踪方法、基于集成学习的深度目标跟踪方法和基于元学习的深度目标跟踪方法等;之后,介绍了目前主要的适用于深度目标跟踪的数据库及其评测方法;接下来从移动端跟踪系统,基于检测与跟踪的系统等方面深入分析与总结了目标跟踪中的最新具体应用情况,最后对深度学习方法在目标跟踪中存在的训练数据不足、实时跟踪和长程跟踪等问题进行分析,并对未来的发展方向进行了展望。     

Discriminative subspace learning with sparse representation view-based model for robust visual tracking

In this paper, we propose a robust tracking algorithm to handle drifting problem. This algorithm consists of two parts: the first part is the G&D part that combines Generative model and Discriminative model for tracking, and the second part is the View-Based model for target appearance that corrects the result of the G&D part if necessary. In G&D part, we use the Maximum Margin Projection (MMP) to construct a graph model to preserve both local geometrical and discriminant structures of the data manifold in low dimensions. Therefore, such discriminative subspace combined with traditional generative subspace can benefit from both models. In addition, we address the problem of learning maximum margin projection under the Spectral Regression (SR) which results in significant savings in computational time. To further solve the drift, an online learned sparsely represented view-based model of the target is complementary to the G&D part. When the result of G&D part is unreliable, the view-based model can rectify the result in order to avoid drifting. Experimental results on several challenging video sequences demonstrate the effectiveness and robustness of our approach.     

Adaptable deep learning structures for object labeling/tracking under dynamic visual environments

Multimedia Tools and Applications - In this paper, we propose a self-adaptive deep neural network architecture suitable for object tracking and labelling. In particular, an adaptation mechanism is...     

Learning structured visual dictionary for object tracking

In this paper, we propose a visual tracking algorithm by incorporating the appearance information gathered from two collaborative feature sets and exploiting its geometric structures. A structured visual dictionary (SVD) can be learned from both appearance and geometric structure, thereby enhancing its discriminative strength between the foreground object and the background. Experimental results show that the proposed tracking algorithm using SVD (SVDTrack) performs favorably against the state-of-the-art methods.     

特征融合与视觉目标跟踪 *

针对跟踪过程中各类图像特征分离背景和目标能力的变化 ,提出一种基于增量判别分析的特征融合算法。该算法首先计算各特征图像的似然图 ,然后通过增量判别分析计算各特征分类性能 ,得到相应权重 ,并在此基础上求取融合似然图 ,通过粒子滤波算法确定待跟踪目标状态。通过对可见光及红外成像视频序列的仿真表明,该算法对环境光照变化、视角变化以及局部遮挡等均具有一定的鲁棒性。     

子空间描述的关节式目标跟踪

针对关节式目标变化对子空间描述造成的影响,本文提出了一种基于增量学习的关节式目标跟踪算法.该算法通过引入图像分割方法与快速傅里叶变换可有效消除背景像素对目标描述造成的影响以及目标区域前景目标位置对不准造成的误差,同时应用局部二值模式增加目标描述中像素点间的几何位置信息,应用基于增量学习的方法实现目标特征的在线更新,最终为跟踪算法提供较为精确的目标描述.实验结果表明,本文提出的关节式目标跟踪算法具有较好的目标跟踪效果.     

Multi-task non-negative matrix factorization for visual object tracking

Pattern Analysis and Applications - This paper proposes an online object tracking algorithm in which the object tracking is achieved by using multi-task sparse learning and non-negative matrix...     

A visual attention model for robot object tracking

Inspired by human behaviors, a robot object tracking model is proposed on the basis of visual attention mechanism, which is fit for the theory of topological perception. The model integrates the image-driven, bottom-up attention and the object-driven, top-down attention, whereas the previous attention model has mostly focused on either the bottom-up or top-down attention. By the bottom-up component, the whole scene is segmented into the ground region and the salient regions. Guided by top-down strategy which is achieved by a topological graph, the object regions are separated from the salient regions. The salient regions except the object regions are the barrier regions. In order to estimate the model, a mobile robot platform is developed, on which some experiments are implemented. The experimental results indicate that processing an image with a resolution of 752*480 pixels takes less than 200ms and the object regions are unabridged. The analysis obtained by comparing the proposed model with the existing model demonstrates that the proposed model has some advantages in robot object tracking in terms of speed and efficiency.     

视觉目标跟踪方法研究综述

目的 随着军事侦察任务设备的发展,红外与可见光侦察技术成为军事装备中的主要侦察手段。研究视觉目标跟踪技术对提高任务设备的全天候目标侦察、目标跟踪、目标定位等战场情报获取能力具有重要意义。目前,对视觉目标跟踪技术的研究越来越深入,目标跟踪的方法和种类也越来越丰富。本文对目前应用较为广泛的4种视觉目标跟踪方法进行研究综述,为后续国内外研究者对目标跟踪相关理论及发展研究工作提供基础。方法 通过对视觉目标跟踪技术难点问题进行分析,根据目标跟踪方法建模方式的不同,将视觉目标跟踪方法分为生成式模型方法与判别式模型方法。分别对生成式模型跟踪算法中的均值漂移目标跟踪方法和粒子滤波目标跟踪方法,判别式模型跟踪算法中的相关滤波目标跟踪方法和深度学习目标跟踪方法进行研究。首先分别对4种跟踪算法的基本原理进行介绍,然后针对4种跟踪算法基本原理的不足和对应目标跟踪中的难点问题进行分析,最后针对目标跟踪的难点问题,给出对应算法的主流改进方案。结果 针对视觉目标跟踪相关技术研究进展,结合无人机侦察任务需求,对跟踪算法实际应用中存在的重点解决问题与相关目标跟踪的难点问题进行分析,给出目前的解决方案与不足,探讨研究未来无人机目标侦察跟踪技术的发展方向。结论 视觉目标跟踪技术已经取得了显著的进展,在侦察任务中的应用越来越广泛。但目标跟踪技术仍然是非常具有挑战性的问题,目标跟踪中的相关理论有待进一步完善和改进,由于实际应用中的场景复杂,目标跟踪的难点问题的挑战性更大,因此容易导致跟踪效果不佳。针对不同的应用环境,结合具体不同军事装备的特点,研究相对精确和鲁棒并且满足实时性要求的视觉目标跟踪算法,对提升装备的全天候侦察目标信息获取能力具有重要意义。     

利用时空特征编码的单目标跟踪网络

目的 随着深度神经网络的出现,视觉跟踪快速发展,视觉跟踪任务中的视频时空特性,尤其是时序外观一致性(temporal appearance consistency)具有巨大探索空间。本文提出一种新颖简单实用的跟踪算法——时间感知网络(temporal-aware network, TAN),从视频角度出发,对序列的时间特征和空间特征同时编码。方法 TAN内部嵌入了一个新的时间聚合模块(temporal aggregation module, TAM)用来交换和融合多个历史帧的信息,无需任何模型更新策略也能适应目标的外观变化,如形变、旋转等。为了构建简单实用的跟踪算法框架,设计了一种目标估计策略,通过检测目标的4个角点,由对角构成两组候选框,结合目标框选择策略确定最终目标位置,能够有效应对遮挡等困难。通过离线训练,在没有任何模型更新的情况下,本文提出的跟踪器TAN通过完全前向推理(fully feed-forward)实现跟踪。结果 在OTB(online object tracking:a benchmark)50、OTB100、TrackingNet、LaSOT(a high-qua...     

Illumination invariant object tracking with adaptive sparse representation

Since the introduction of the sparse representation-based tracking method named ?₁ tracker, there have been further studies into this tracking framework with promised results in challenging video sequences. However, in the situation of large illumination changes and shadow casting, the tracked object cannot be modeled efficiently by sparse representation templates. To overcome this problem, we propose a new illumination invariant tracker based on photometric normalization techniques and the sparse representation framework. With photometric normalization methods, we designed a new illumination invariant template presentation for tracking that eliminates the illumination influences, such as brightness variation and shadow casting. For a higher tracking accuracy, we introduced a strategy that adaptively selects the optimum template presentation at the update step of the tracking process. The experiments show that our approach outperforms the previous ?₁ and some state-of-the-art algorithms in tracking sequences with severe illumination effects.     

SiamMN: Siamese modulation network for visual object tracking

Multimedia Tools and Applications - Visual object tracking methods based on Siamese network are often difficult to distinguish objects with the same semantic or similar appearance as tracking...     

Online unsupervised feature learning for visual tracking

We propose a method for visual tracking-by-detection based on online feature learning. Our learning framework performs feature encoding with respect to an over-complete dictionary, followed by spatial pyramid pooling. We then learn a linear classifier based on the resulting feature encoding. Unlike previous work, we learn the dictionary online and update it to help capture the appearance of the tracked target as well as the background. In more detail, given a test image window, we extract local image patches from it and each local patch is encoded with respect to the dictionary. The encoded features are then pooled over a spatial pyramid to form an aggregated feature vector. Finally, a simple linear classifier is trained on these features.Our experiments show that the proposed powerful—albeit simple—tracker, outperforms all the state-of-the-art tracking methods that we have tested. Moreover, we evaluate the performance of different dictionary learning and feature encoding methods in the proposed tracking framework, and analyze the impact of each component in the tracking scenario. In particular, we show that a small dictionary, learned and updated online is as effective and more efficient than a huge dictionary learned offline. We further demonstrate the flexibility of feature learning by showing how it can be used within a structured learning tracking framework. The outcome is one of the best trackers reported to date, which facilitates the advantages of both feature learning and structured output prediction. We also implement a multi-object tracker, which achieves state-of-the-art performance.