Adaptive bootstrapping management by keypoint clustering for background initialization

Detection of video objects under bad weather and poor illumination condition is a challenging task. We address this issue using the notion of background modeling. LBP-based background modeling and model learning has been used to detect the video objects but performance degrades in the above complex background. We propose the adjacency- and reinforced adjacency-based variants of LBP for complex real-world background modeling and model learning for object detection. In this regard, we have proposed the four variants; (1) enhanced adjacent local binary pattern, (2) enhanced reinforced adjacent local binary pattern, (3) enhanced adjacent local ternary pattern, and (4) enhanced reinforced adjacent local ternary pattern. Besides, we have embedded the Gabor and LBP features to obtain an embedded feature, which is subsequently used with the notions of adjacency. Unlike the background learning approach where the model learns only the background, our model learning algorithm learns the background together with the foreground objects and hence named as mixed-mode learning strategy. These models together with the new learning strategy are tested with CD2014 (snowfall and blizzard) and PETS (2014 and 2016) data sets, and the performance of the proposed models has been compared with LBP-based methods and other state-of-the-art methods.

We investigate how hybrid camera systems---stationary and mobile cameras---allow to improve the observability of mobile objects and/or locations considering limited resources. This Static Camera System with Mobile robots (SCSM) allows mobile cameras to observe targets where no stationary cameras are deployed, where they fail, or their field of view (FOV) is blocked. By assigning dynamic priorities to targets, the self-coordination and control of the SCSM becomes an assignment problem. Furthermore, the coordination of the available resources, especially for the mobile part of the SCSM, is considered. We adapt a market-based approach based on dynamic clustering. The SCSM is evaluated through simulation studies with a graphical simulator for visual sensor networks (VSNs). The results show that the SCSM can achieve up to 30% higher observability of targets compared to a stationary camera system. Moreover, the resource consumption can be distributed among the cameras with the dynamic clustering protocol to not burden all cameras having the same object in their FOV.

Background subtraction is an important part of various computer vision applications that can detect the foreground objects by comparing the current pixels with a background model. The general approaches gradually update the background model according to the current status, but might fail in sudden illumination changes. An illumination‐robust background modelling method is proposed to address this problem. The method provides quick illumination compensation using two background models with different adaption rates. Experimental results show that the proposed algorithm outperforms several state‐of‐art approaches and provides low computational cost.

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.

This paper presents a hybrid pixel-based background (HPB) model, which is constructed by single stable record and multi-layer astable records after initial learning. The image foreground object detection must face the problems of moving background, illumination changes, chaotic, etc. in real word applications. In our approach, the HPB model can be used for background subtraction to extract objects precisely in various complex scenes. Using the multi-layer astable records, we also propose the homogeneous background subtraction that can detect the foreground object with less record memory. Based on the benchmark videos, the experimental results show that single stable and 3-layer multi-layer astable records can be enough for background model construction and then updated quickly to overcome the background variation. The proposed approach can improve the averages Error Rate of foreground object detection up to 86% when comparing with the latest works.

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] Camera traps are a popular tool to sample animal populations because they are noninvasive, detect a variety of species, and can record many thousands of animal detections per deployment. Cameras are typically set to take bursts of multiple images for each detection, and are deployed in arrays of dozens or hundreds of sites, often resulting in millions of images per study. The task of converting images to animal detection records from such large image collections is daunting, and made worse by situations that generate copious empty pictures from false triggers (e.g. camera malfunction or moving vegetation) or pictures of humans. We offer the first widely available computer vision tool for processing camera trap images. Our results show that the tool is accurate and results in substantial time savings for processing large image datasets, thus improving our ability to monitor wildlife across large scales with camera traps. In this dissertation, we have developed new image/video processing and computer vision algorithms for efficient and accurate object detection and sequence-level classiffication from natural scene camera-trap images. This work addresses the following five major tasks: (1) Human-animal detection. We develop a fast and accurate scheme for human-animal detection from highly cluttered camera-trap images using joint background modeling and deep learning classification. Specifically, first, We develop an effective background modeling and subtraction scheme to generate region proposals for the foreground objects. We then develop a cross-frame image patch verification to reduce the number of foreground object proposals. Finally, We perform complexity-accuracy analysis of deep convolutional neural networks (DCNN) to develop a fast deep learning classification scheme to classify these region proposals into three categories: human, animals, and background patches. The optimized DCNN is able to maintain high level of accuracy while reducing the computational complexity by 14 times. Our experimental results demonstrate that the proposed method outperforms existing methods on the camera-trap dataset. (2) Object segmentation from natural scene. We first design and train a fast DCNN for animal-human-background object classification, which is used to analyze the input image to generate multi-layer feature maps, representing the responses of different image regions to the animal-human-background classifier. From these feature maps, we construct the so-called deep objectness graph for accurate animal-human object segmentation with graph cut. The segmented object regions from each image in the sequence are then verfied and fused in the temporal domain using background modeling. Our experimental results demonstrate that our proposed method outperforms existing state-of-the-art methods on the camera-trap dataset with highly cluttered natural scenes. (3) DCNN domain background modeling. We replaced the background model with a new more efficient deep learning based model. The input frames are segmented into regions through the deep objectness graph then the region boundaries of the input frames are multiplied by each other to obtain the regions of movement patches. We construct the background representation using the temporal information of the co-located patches. We propose to fuse the subtraction and foreground/background pixel classiffcation of two representation : a) chromaticity and b) deep pixel information. (4) Sequence-level object classiffcation. We proposed a new method for sequence-level video recognition with application to animal species recognition from camera trap images. First, using background modeling and cross-frame patch verification, we developed a scheme to generate candidate object regions or object proposals in the spatiotemporal domain. Second, we develop a dynamic programming optimization approach to identify the best temporal subset of object proposals. Third, we aggregate and fuse the features of these selected object proposals for efficient sequence-level animal species classification.

In its early implementations, background modeling was a process of building a model for the background of a video with a stationary camera, and identifying pixels that did not conform well to this model. The pixels that were not well-described by the background model were assumed to be moving objects. Many systems today maintain models for the foreground as well as the background, and these models compete to explain the pixels in a video. If the foreground model explains the pixels better, they are considered foreground. Otherwise they are considered background. In this paper, we argue that the logical endpoint of this evolution is to simply use Bayes' rule to classify pixels. In particular, it is essential to have a background likelihood, a foreground likelihood, and a prior at each pixel. A simple application of Bayes' rule then gives a posterior probability over the label. The only remaining question is the quality of the component models: the background likelihood, the foreground likelihood, and the prior. We describe a model for the likelihoods that is built by using not only the past observations at a given pixel location, but by also including observations in a spatial neighborhood around the location. This enables us to model the influence between neighboring pixels and is an improvement over earlier pixelwise models that do not allow for such influence. Although similar in spirit to the joint domain-range model, we show that our model overcomes certain deficiencies in that model. We use a spatially dependent prior for the background and foreground. The background and foreground labels from the previous frame, after spatial smoothing to account for movement of objects, are used to build the prior for the current frame. These components are, by themselves, not novel aspects in background modeling. As we will show, many existing systems account for these aspects in different ways. We argue that separating these components as suggested in this paper yields a very simple and effective model. Our intuitive description also isolates the model components from the classification or inference step. Improvements to each model component can be carried out without any changes to the inference or other components. The various components can hence be modeled effectively and their impact on the overall system understood more easily.

We developed an automatic object segmentation system to separate the foreground objects from the background scene in the MPEG video sequence. The system consists of two modules: the background modeling and updating module and the foreground object extraction module. For the first module and comparing to existing methods, the background model can be constructed no matter whether there exist moving foreground objects or not. In addition, the background model is capable of handling the illumination changes and intrusive but motionless targets by using the short-term approach and long-term approach, respectively, to keep updating the background model. For the second module, the noises and shadows are eliminated and the holes are filled in order to reduce the false and the missing foreground detection components, respectively. Furthermore, one particular function in this module is the automatic digital matting, which can be applied to have visually accurate segmentation result for the foreground objects.KeywordsVideo SequenceBackground SubtractionBackground ModelForeground ObjectObject SegmentationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Background subtraction is a widely used technique for detecting moving objects in image sequences. Very often background subtraction approaches assume the availability of one or more clear frames (i.e., without foreground objects) at the beginning of the image sequence in input. This strong assumption is not always correct, especially when dealing with dynamic background. In this paper, we present the results of an on-line and real-time background initialization method, called IMBS, which generates a reliable initial background model even if no clear frames are available. The accuracy of the proposed approach is calculated on a set of seven publicly available benchmark sequences. Experimental results demonstrate that IMBS generates accurate background models with respect to eight different quality metrics.

ViBe is a pixel-level detection algorithm. However, the algorithm is easy to introduce the ghost region by establishing the initial background model in the first frame of the image, and the fixed threshold for background update cannot adapt to the scene with large background changes. The above problems may cause the error detection of the moving target. This paper presents an improved algorithm to initialize the background model using multi-frame image pixels. In the detection of moving target, the adaptive threshold is set according to the distance between the current pixel value and the background model and the measured value of frame difference. Create a counter for each foreground point to quickly eliminate ghosts. Experimental results show that the improved algorithm can suppress the ghost more effectively and detect the foreground target more accurately.

Background subtraction in real applications: Challenges, current models and future directions

This paper presents a real-time surveillance system for detecting and tracking people, which takes full advantage of local texture patterns, under a stationary monocular camera. A novel center-symmetric scale invariant local ternary pattern feature is put forward to combine with pattern kernel density estimation for building a pixel-level-based background model. The background model is then used to detect moving foreground objects on every newly captured frame. A variant of a fast human detector that utilizes local texture patterns is adopted to look for human objects from the foreground regions, and it is assisted by a head detector, which is proposed to find in advance the candidate locations of human, to reduce computational costs. Each human object is given a unique identity and is represented by a spatio-color-texture object model. The real-time performance of tracking is achieved by a fast mean-shift algorithm coupled with several efficient occlusion-handling techniques. Experiments on challenging video sequences show that the proposed surveillance system can run in real-time and is quite robust in segmenting and tracking people in complex environments that include appearance changes, abrupt motion, occlusions, illumination variations and clutter.

This paper describes a new background subtraction method for detecting moving foreground objects in video from a static camera. The background model assumes that for each frame, every pixel is described as either an image patch, or a texture patch. An image patch describes the local appearance of an image, and is suitable when the background is structured and of fixed appearance. A texture patch is useful for describing a pixel which is part of a larger region containing a consistent patterning of pixels, but for which the local appearance is either indistinct (e.g. uniform regions), or time varying (e.g. the surface of the sea). Textures tend to correspond to regions rather than specific locations, so the texture descriptions are calculated globally (i.e. over the entire image). Over time, each pixel is represented as a weighted mixture of modes, as in the Stauffer and Grimson method. But instead of each mode being an image intensity with a variance, each mode now corresponds to either an image or a texture patch. The proposed background model is applied to the nine video sequences from the I2R data set, and compared to the provided foreground ground-truth. When the error is measured on a per pixel basis, the proposed method performs better than the alternative methods tested.

One of the fundamental operations in computer vision applications is change detection, in which moving foreground objects are segmented from a static background. A common approach for change detection is the comparison of an image frame with the stored background model using a matching algorithm, a process known as background subtraction. However, such techniques fail in environments with dynamic backgrounds, illumination changes, or shadow and camera jitters. This study focuses on effectively detecting scene changes in complex environments. To this end, we proposed a new colour descriptor named local colour difference pattern (LCDP) that is insusceptible to shadow and is able to capture both colour and texture features at a pixel location. Furthermore, a scene change detection framework was proposed to handle dynamic scenes based on sample consensus that integrates LCDP and a novel spatial model fusion mechanism. Experiments using the CDnet benchmark dataset demonstrated the effectiveness of the proposed approach to change detection in complex environments.

We propose a new method for background modeling. Our method is based on the two complementary approaches. One uses the probability density function(PDF) to approximate background model. The PDF is estimated non-parametrically by using Parzen density estimation. And foreground object is detected based on the estimated PDF. The other method is based on the evaluation of the local texture at pixel-level resolution while reducing the effects of variations in lighting. Fusing their approach realize robust object detection under varying illumination. Several experiments show the effectiveness of our approach.