Sequential kernel density approximation and its application to real-time visual tracking

Analysis of spatial structure and influencing factors of the distribution of national industrial heritage sites in China based on mathematical calculations

An in-depth analysis of the spatial distribution characteristics and overall pattern of industrial heritage sites in China provides not only a comprehensive understanding of the current status of industrial heritage but also a reference for its protection and ongoing utilization. A total of 170 industrial heritage sites that were included in the List of National Industrial Heritage of China were selected as the research objects. Their spatial structure characteristics were quantitatively analyzed based on a kernel density analysis of ArcGIS and imbalance and Gini coefficient index of function calculations. The results show that the distribution of industrial heritage sites in China presents a strong aggregation trend and a distribution pattern of four cores, six centers, and multiple scattered points. The distribution of industrial heritage sites in 34 administrative regions is extremely imbalanced. A total of 170 industrial heritage sites are distributed across 27 administrative regions; 52.35% are concentrated in the East and Southwest divisions. According to the index definitions, this research analyzed their influencing factors from perspectives of the natural and social environments. The results show that the industrial heritage sites in China are mainly distributed in traditional agricultural and commercial areas with rich natural or water transport resources. The current study of major historical events in modern China and the growth curve of industrial heritage concludes that China's industry has experienced five stages of development: Ancient, Beginning, Accelerated Development, Climax, and Slowdown. The geographical divisions and distribution of categories show colonial and socialist characteristics.

Brain Metastases From Lung Adenocarcinoma May Preferentially Involve the Distal Middle Cerebral Artery Territory and Cerebellum

Although whole-brain radiation therapy (WBRT) is the mainstay of treatment for brain metastases (BMs), the concept of saving eloquent cortical lesions has been promoted. If BMs from lung cancer are spatially biased to certain regions, this approach can be justified more. We evaluated whether BMs from lung cancer show a preference for certain brain regions and if their distribution pattern differs according to the histologic subtype of the primary lung cancer. In this retrospective study, 562 BMs in 80 patients were analyzed (107 BMs from small cell carcinoma, 432 from adenocarcinoma, and 23 from squamous cell carcinoma). Kernel density estimation was performed to investigate whether BM spatial patterns differed among lung cancer subtypes. Further, we explored more detailed subregions where BMs from adenocarcinomas occur frequently using one-way analysis of variance. Finally, we divided our cohort into those with fewer (≤10) and more (>10) BMs and evaluated whether this biased pattern was maintained across limited and extensive stages. For small cell carcinoma, BMs were biased to the cerebellum, but this did not reach statistical significance. For adenocarcinoma, BMs were found more frequently near the distal middle cerebral artery (MCA) territory and cerebellum than in other arterial territories (p < 0.01). The precentral and postcentral gyri were the most significant subregions within the distal anterior cerebral artery (ACA) and MCA territories (p < 0.01). Crus I and Lobule VI were significant regions within the cerebellum (p < 0.01). Regardless of the number of BMs, the affinity to the distal MCA territory and cerebellum was maintained. The present data confirm that BMs from lung adenocarcinoma may preferentially involve the distal MCA territory and cerebellum.

Extended motion diffusion based change detection for airport ground surveillance

Change detection in airport ground is important for airport security. Due to the particularity of ground environment, e.g. haze and camouflage, airport ground change detection is generally incomplete. If an incomplete detection is used as reference for the detection in subsequent frames, it may result in noticeable detection defects across the frames. In this paper, extended motion diffusion (EMD) is proposed to address the problems. The core idea of the EMD is to design a novel model insensitive to incomplete detection. Firstly the one-to-many correspondence in traditional motion diffusion is extended in the prediction step of EMD to build up correspondence from incomplete detection to intact objects. Prior information, e.g. aircraft motion prior and ground structure prior, is employed in the development of the correspondence. Then based on the correspondence a number of new samples are synthesized and filtered in the identification step of the EMD to compensate possible detection defects. Finally, the reserved samples are collected to train a foreground model, which is used in conjunction with another background model for classification. The proposed method is verified based on the Airport Ground Video Surveillance (AGVS) benchmark. Experimental results show effectiveness of the proposed algorithm in dealing with haze and camouflage.

A Fusion Framework for Camouflaged Moving Foreground Detection in the Wavelet Domain

Detecting camouflaged moving foreground objects has been known to be difficult due to the similarity between the foreground objects and the background. Conventional methods cannot distinguish the foreground from background due to the small differences between them and thus suffer from underdetection of the camouflaged foreground objects. In this paper, we present a fusion framework to address this problem in the wavelet domain. We first show that the small differences in the image domain can be highlighted in certain wavelet bands. Then the likelihood of each wavelet coefficient being foreground is estimated by formulating foreground and background models for each wavelet band. The proposed framework effectively aggregates the likelihoods from different wavelet bands based on the characteristics of the wavelet transform. Experimental results demonstrated that the proposed method significantly outperformed existing methods in detecting camouflaged foreground objects. Specifically, the average F-measure for the proposed algorithm was 0.87, compared to 0.71 to 0.8 for the other stateof- the-art methods.

Learning based particle filtering object tracking for visible-light systems

We propose a novel object tracking framework based on online learning scheme that can work robustly in challenging scenarios. Firstly, a learning-based particle filter is proposed with color and edge-based features. We train a. support vector machine (SVM) classifier with object and background information and map the outputs into probabilities, then the weight of particles in a particle filter can be calculated by the probabilistic outputs to estimate the state of the object. Secondly, the tracking loop starts with Lucas-Kanade (LK) affine template matching and follows by learning-based particle filter tracking. Lucas-Kanade method estimates errors and updates object template in the positive samples dataset, and learning-based particle filter tracker will start if the LK tracker loses the object. Finally, SVM classifier evaluates every tracked appearance to update the training set or restart the tracking loop if necessary. Experimental results show that our method is robust to challenging light, scale and pose changing, and test on eButton image sequence also achieves satisfactory tracking performance.

Online Discriminative Kernel Density Estimator With Gaussian Kernels

We propose a new method for a supervised online estimation of probabilistic discriminative models for classification tasks. The method estimates the class distributions from a stream of data in the form of Gaussian mixture models (GMMs). The reconstructive updates of the distributions are based on the recently proposed online kernel density estimator (oKDE). We maintain the number of components in the model low by compressing the GMMs from time to time. We propose a new cost function that measures loss of interclass discrimination during compression, thus guiding the compression toward simpler models that still retain discriminative properties. The resulting classifier thus independently updates the GMM of each class, but these GMMs interact during their compression through the proposed cost function. We call the proposed method the online discriminative kernel density estimator (odKDE). We compare the odKDE to oKDE, batch state-of-the-art kernel density estimators (KDEs), and batch/incremental support vector machines (SVM) on the publicly available datasets. The odKDE achieves comparable classification performance to that of best batch KDEs and SVM, while allowing online adaptation from large datasets, and produces models of lower complexity than the oKDE.

A survey of appearance models in visual object tracking

Visual object tracking is a significant computer vision task which can be applied to many domains, such as visual surveillance, human computer interaction, and video compression. Despite extensive research on this topic, it still suffers from difficulties in handling complex object appearance changes caused by factors such as illumination variation, partial occlusion, shape deformation, and camera motion. Therefore, effective modeling of the 2D appearance of tracked objects is a key issue for the success of a visual tracker. In the literature, researchers have proposed a variety of 2D appearance models.

To help readers swiftly learn the recent advances in 2D appearance models for visual object tracking, we contribute this survey, which provides a detailed review of the existing 2D appearance models. In particular, this survey takes a module-based architecture that enables readers to easily grasp the key points of visual object tracking. In this survey, we first decompose the problem of appearance modeling into two different processing stages: visual representation and statistical modeling. Then, different 2D appearance models are categorized and discussed with respect to their composition modules. Finally, we address several issues of interest as well as the remaining challenges for future research on this topic.

The contributions of this survey are fourfold. First, we review the literature of visual representations according to their feature-construction mechanisms (i.e., local and global). Second, the existing statistical modeling schemes for tracking-by-detection are reviewed according to their model-construction mechanisms: generative, discriminative, and hybrid generative-discriminative. Third, each type of visual representations or statistical modeling techniques is analyzed and discussed from a theoretical or practical viewpoint. Fourth, the existing benchmark resources (e.g., source codes and video datasets) are examined in this survey.

Trajectory data analyses for pedestrian space-time activity study

It is well recognized that human movement in the spatial and temporal dimensions has direct influence on disease transmission(1-3). An infectious disease typically spreads via contact between infected and susceptible individuals in their overlapped activity spaces. Therefore, daily mobility-activity information can be used as an indicator to measure exposures to risk factors of infection. However, a major difficulty and thus the reason for paucity of studies of infectious disease transmission at the micro scale arise from the lack of detailed individual mobility data. Previously in transportation and tourism research detailed space-time activity data often relied on the time-space diary technique, which requires subjects to actively record their activities in time and space. This is highly demanding for the participants and collaboration from the participants greatly affects the quality of data(4). Modern technologies such as GPS and mobile communications have made possible the automatic collection of trajectory data. The data collected, however, is not ideal for modeling human space-time activities, limited by the accuracies of existing devices. There is also no readily available tool for efficient processing of the data for human behavior study. We present here a suite of methods and an integrated ArcGIS desktop-based visual interface for the pre-processing and spatiotemporal analyses of trajectory data. We provide examples of how such processing may be used to model human space-time activities, especially with error-rich pedestrian trajectory data, that could be useful in public health studies such as infectious disease transmission modeling. The procedure presented includes pre-processing, trajectory segmentation, activity space characterization, density estimation and visualization, and a few other exploratory analysis methods. Pre-processing is the cleaning of noisy raw trajectory data. We introduce an interactive visual pre-processing interface as well as an automatic module. Trajectory segmentation(5) involves the identification of indoor and outdoor parts from pre-processed space-time tracks. Again, both interactive visual segmentation and automatic segmentation are supported. Segmented space-time tracks are then analyzed to derive characteristics of one's activity space such as activity radius etc. Density estimation and visualization are used to examine large amount of trajectory data to model hot spots and interactions. We demonstrate both density surface mapping(6) and density volume rendering(7). We also include a couple of other exploratory data analyses (EDA) and visualizations tools, such as Google Earth animation support and connection analysis. The suite of analytical as well as visual methods presented in this paper may be applied to any trajectory data for space-time activity studies.

A novel recursive Bayesian learning-based method for the efficient and accurate segmentation of video with dynamic background

Segmentation of video with dynamic background is an important research topic in image analysis and computer vision domains. In this paper, we present a novel recursive Bayesian learning-based method for the efficient and accurate segmentation of video with dynamic background. In the algorithm, each frame pixel is represented as the layered normal distributions which correspond to different background contents in the scene. The layers are associated with a confident term and only the layers satisfy the given confidence which will be updated via the recursive Bayesian estimation. This makes learning of background motion trajectories more accurate and efficient. To improve the segmentation quality, the coarse foreground is obtained via simple background subtraction first. Then, a local texture correlation operator is introduced to fill the vacancies and remove the fractional false foreground regions. Extensive experiments on a variety of public video datasets and comparisons with some classical and recent algorithms are used to demonstrate its improvements in both segmentation accuracy and efficiency.

Multistage particle windows for fast and accurate object detection

The common paradigm employed for object detection is the sliding window (SW) search. This approach generates grid-distributed patches, at all possible positions and sizes, which are evaluated by a binary classifier: The tradeoff between computational burden and detection accuracy is the real critical point of sliding windows; several methods have been proposed to speed up the search such as adding complementary features. We propose a paradigm that differs from any previous approach since it casts object detection into a statistical-based search using a Monte Carlo sampling for estimating the likelihood density function with Gaussian kernels. The estimation relies on a multistage strategy where the proposal distribution is progressively refined by taking into account the feedback of the classifiers. The method can be easily plugged into a Bayesian-recursive framework to exploit the temporal coherency of the target objects in videos. Several tests on pedestrian and face detection, both on images and videos, with different types of classifiers (cascade of boosted classifiers, soft cascades, and SVM) and features (covariance matrices, Haar-like features, integral channel features, and histogram of oriented gradients) demonstrate that the proposed method provides higher detection rates and accuracy as well as a lower computational burden w.r.t. sliding window detection.

Density-based multifeature background subtraction with support vector machine

Background modeling and subtraction is a natural technique for object detection in videos captured by a static camera, and also a critical preprocessing step in various high-level computer vision applications. However, there have not been many studies concerning useful features and binary segmentation algorithms for this problem. We propose a pixelwise background modeling and subtraction technique using multiple features, where generative and discriminative techniques are combined for classification. In our algorithm, color, gradient, and Haar-like features are integrated to handle spatio-temporal variations for each pixel. A pixelwise generative background model is obtained for each feature efficiently and effectively by Kernel Density Approximation (KDA). Background subtraction is performed in a discriminative manner using a Support Vector Machine (SVM) over background likelihood vectors for a set of features. The proposed algorithm is robust to shadow, illumination changes, spatial variations of background. We compare the performance of the algorithm with other density-based methods using several different feature combinations and modeling techniques, both quantitatively and qualitatively.

Shared Kernel Information Embedding for discriminative inference

Latent variable models, such as the GPLVM and related methods, help mitigate overfitting when learning from small or moderately sized training sets. Nevertheless, existing methods suffer from several problems: 1) complexity, 2) the lack of explicit mappings to and from the latent space, 3) an inability to cope with multimodality, and 4) the lack of a well-defined density over the latent space. We propose an LVM called the Kernel Information Embedding (KIE) that defines a coherent joint density over the input and a learned latent space. Learning is quadratic, and it works well on small data sets. We also introduce a generalization, the shared KIE (sKIE), that allows us to model multiple input spaces (e.g., image features and poses) using a single, shared latent representation. KIE and sKIE permit missing data during inference and partially labeled data during learning. We show that with data sets too large to learn a coherent global model, one can use the sKIE to learn local online models. We use sKIE for human pose inference.

Recognizing human actions by learning and matching shape-motion prototype trees

A shape-motion prototype-based approach is introduced for action recognition. The approach represents an action as a sequence of prototypes for efficient and flexible action matching in long video sequences. During training, an action prototype tree is learned in a joint shape and motion space via hierarchical K-means clustering and each training sequence is represented as a labeled prototype sequence; then a look-up table of prototype-to-prototype distances is generated. During testing, based on a joint probability model of the actor location and action prototype, the actor is tracked while a frame-to-prototype correspondence is established by maximizing the joint probability, which is efficiently performed by searching the learned prototype tree; then actions are recognized using dynamic prototype sequence matching. Distance measures used for sequence matching are rapidly obtained by look-up table indexing, which is an order of magnitude faster than brute-force computation of frame-to-frame distances. Our approach enables robust action matching in challenging situations (such as moving cameras, dynamic backgrounds) and allows automatic alignment of action sequences. Experimental results demonstrate that our approach achieves recognition rates of 92.86 percent on a large gesture data set (with dynamic backgrounds), 100 percent on the Weizmann action data set, 95.77 percent on the KTH action data set, 88 percent on the UCF sports data set, and 87.27 percent on the CMU action data set.