Background Subtraction Model Research Articles

Efficient local monitoring approach for the task of background subtraction

We present in this paper a novel and efficient method that will significantly reduce GMM drawbacks in the presence of complex and dynamic scene. The main idea is to combine global and local features to remove local variations and the instant variations in the brightness that, in most cases, decrease the performance of background subtraction models. The first step is to divide the extracted frames into several equal size blocks. Then, we apply an adaptive local monitoring algorithm for each block to control local variation using Pearson similarity measurement. When a significant environment changes are detected in one or more blocks, the parameters of GMM assigned to these blocks are updated and the parameters of the rest remain the same. We also proposed merging adjacent and invariant blocks to reduce processing time and splitting the blocks that have an intense movement to improve accuracy. Experimental results on several datasets demonstrate that the proposed approach is effective and efficient under the common problems found in background modeling, outperforming the most referred state-of-the-art background subtraction methods.

An edge-based method for effective abandoned luggage detection in complex surveillance videos

Abandoned objects detection is one of the most challenging tasks in intelligent video surveillance systems. In this paper we present a new method for detecting abandoned objects (AO) using edges instead of pixel intensities. Our main focus, is on reducing false alarms, while keeping a high positives detection rate. Based on edge information, the proposed method reduces errors rate considerably compared to pixel intensities based approaches. At first, static edges are detected by applying a temporal accumulation step using the foreground edges mask resulting from the edge-based background subtraction model. Then, edges clustering is applied on the obtained stable edges mask, using edges’ position and stability in time to delineate the object bounding box. Finally, an efficient classification approach, relying on edges’ position, orientation, and staticness based scores, is applied on the AO candidates. The proposed approach has been validated on several challenging benchmarks, and is compared to other works of the literature. The results shows that our method reduce false alarms rates greatly, while keeping a good detection accuracy of true positives.

Camera movement artefact correction

Traffic surveillance systems based on background subtraction model are vulnerable to movements or vibrations in camera posts. Such camera shaking reflects in the form of relative displacements in locations of stationary structures and misinterpreted as valid motion. This artefact is referred to as camera movement artefact. The proposed method extracts the possible regions vulnerable to false motion detection and creates a mask which subtracts out the artefact. The real-time system in which the proposed method is implemented, the allowed processing time per frame is limited to < 30 milliseconds, and the artefact correction mechanism is one of the many components. Hence, this simple mask subtraction provides an optimal solution compared to other per-frame correction mechanisms. Performance is evaluated for multiple test sets and the results show significant reduction in rate of false vehicle detection (36% to 12%) without negotiating available correct vehicle detection (98%).

Background subtraction based on circulant matrix

Detecting moving objects in a scene is a fundamental and critical step for many high-level computer vision tasks. However, background subtraction modeling is still an open and challenge problem, particularly in practical scenarios with drastic illumination changes and dynamic backgrounds. In this paper, we present a novel background modeling method focused on dealing with complex environments based on circular shift operator. The background model is constructed by performing circular shifts on the neighborhood of each pixel, which forms a basic region unit. The foreground mask is obtained via two stages. The first stage is to subtract the established background from the current frame to obtain the distance map. The second is to adopt the graph cut on the distance map. In order to adapt to the background changes, the background model is updated with an adaptive update rate. Experimental results on indoor and outdoor videos demonstrate the efficiency of our proposed method.

Vehicle detection in urban traffic scenes using the Pixel-Based Adaptive Segmenter with Confidence Measurement

The Pixel-Based Adaptive Segmenter with Confidence Measurement (PBASCM) is proposed for vehicle detection in complex urban traffic scenes to efficiently address deficiencies of the background subtraction model, which is easily contaminated by slow-moving or temporarily stopped vehicles. The background is modeled based on the history of recently observed pixel values and each pixel in the background model is assigned a confidence measurement based on the current traffic state. The foreground decision depends on an adaptive threshold, whereas the background model is updated based on the current traffic state and whether the corresponding pixel point is in the confidence period. Using real-world urban traffic videos, the overall results of detection accuracy analyses demonstrated that PBASCM achieved better performance in both qualitative and quantitative evaluations, compared with other state-of-the-art methods. PBASCM can accurately detect slowmoving or temporarily stopped vehicles, and the similarity and F-measure results for PBASCM were 0.839 and 0.912 higher, respectively, than those obtained by other methods in a traffic light sequence during the daytime. Thus, our experimental results demonstrate that PBASCM is effective and suitable for real-time vehicle detection in complex urban traffic scenes.

A novel Rotational Symmetry Dynamic Texture (RSDT) based sub space construction and SCD (Similar-Congruent-Dissimilar) based scoring model for background subtraction in real time videos

Background Subtraction (BS) plays an important role in video surveillance system because it provides a focus of attention for moving object detection. But there are still challenging scenarios such as dynamic backgrounds and illumination variation. To achieve greater foreground detection quality under these circumstances, a novel approach called "Rotational Symmetry Dynamic Texture" for Background Subtraction is proposed. The concept of key frame is used to avoid processing overhead. A spatio temporal patch is used to describe motion and appearance of a video sequence. From that patch the proposed Rotational Symmetry Dynamic Texture (RSDT) method will encode the relationship between the referenced pixel and its neighbours, based on the directions that are calculated using the concept of rotational symmetry and line symmetry. A novel "SCD (Similar-Congruent-Dissimilar) based scoring" model is used to estimate appearance consistency and temporal coherence of the video. The two layer approach is used to solve sudden illumination changes and to improve processing speed by adopting mean values of blocks. It is suitable for both indoor and outdoor environments because the subspace construction is purely based on directional codes with rotational symmetry and line symmetry. Metric F-Score is used to measure the performance of the proposed method. Finally this paper provides an effective and robust background subtraction scheme. From performance evaluation, it is observed that Rotational Symmetry Dynamic Texture with SCD based scoring model guarantees accurate background subtraction than the existing literature.

Fast Compressive Target Detection for Wireless Video Sensor Nodes

In order to prolong the lifetime of visual target detection and tracking system based on wireless video sensor networks, many efficient methods have been proposed to reduce the energy consumption of the battery-powered video sensor nodes. Focused on reducing the amount of image data for computing, this paper presents a fast compressive method of target detection for video sensor nodes using structured compressive sensing. The major contributions are as follows: Firstly, we construct a novel structured measurement matrix for sampling the image. Secondly, we use an efficient adaptive Gaussian mixture model for real-time background subtraction. Experimental results show that our method can achieve good performance and over two times faster than traditional Gaussian mixture model.

A Background Foreground Competitive Model for Background Subtraction in Dynamic Background

Background subtraction is used to remove the relatively motionless background information from video frames and to detect moving objects. Several methods have been proposed for nonparametric modelling of the background. In this paper a novel method for simultaneous non parametric modelling of background and foreground is proposed, which is utilized for classification of pixels as foreground or background, in a competitive manner. Selective updating of the background and foreground models is employed to accommodate changes in the background. Both temporal and spatial dependencies of pixels are utilized in the model updating. The proposed method gives higher Percentage of Correct Classification (PCC) score in dynamic background compared to the other methods, which is verified using standard databases.

ViBe with Adaptive Threshold Based on Energy Minimization

VIsual Background Extractor (VIBe) is a simple and effective background subtraction model, it is important to choose the threshold according to different environments. A new method based on the VIBe is presented which uses the adaptive threshold. We give a new description of the similarity degree without threshold, and then use the energy minimization model to combine the spatial information and show that dynamically choosing threshold in our kernel estimates at each individual pixel can significantly improve results. To compare with the traditional VIBe algorithms, a synthetic test SABS and a real world test I2R are used. Experiment shows that the presented method is on average 10% better than the traditional VIBe in accuracy.

A Real-Time Active Pedestrian Tracking System Inspired by the Human Visual System

Pedestrian detection and tracking play a significant role in surveillance. Despite the numerous detection and tracking methods proposed in the literature, when the pedestrian is too small to recognize, which is a common case in modern surveillance systems, all methods fail. In order to deal with such case, we propose an active pedestrian tracking system inspired by the human visual system. A coarse-to-fine pedestrian detection algorithm is proposed for the small pedestrian detection by combining the Gaussian mixture model background subtraction with the histogram of oriented gradient detection. In addition, a three-dimensional pan–tilt–zoom control model is presented, which requires no calibration and is more accurate than other control models. In order to actively track a pedestrian in real time, we utilize an active control algorithm and a tracking–learning–detection tracker. Experimental results demonstrate that our active tracking system is both efficient and effective.

A Self-adaptive CodeBook (SACB) model for real-time background subtraction

Effective and efficient background subtraction is important to a number of computer vision tasks. In this paper, we introduce a new background model that integrates several new techniques to address key challenges for background modeling for moving object detection in videos. The novel features of our proposed Self-adaptive CodeBook (SACB) background model are: a more effective color model using YCbCr color space, a statistical parameter estimation method, and a new algorithm for adding new background codewords into the permanent model and deleting noisy codewords from the models. Also, a new block-based approach is introduced to exploit the local spatial information. The proposed model is rigorously tested and has shown significant performance improvements over several previous models.

Object detection in video sequences by a temporal modular self-adaptive SOM

A video segmentation algorithm that takes advantage of using a background subtraction (BS) model with low learning rate (LLR) or a BS model with high learning rate (HLR) depending on the video scene dynamics is presented in this paper. These BS models are based on a neural network architecture, the self-organized map (SOM), and the algorithm is termed temporal modular self-adaptive SOM, TMSA_SOM. Depending on the type of scenario, the TMSA_SOM automatically classifies and processes each video into one of four different specialized modules based on an initial sequence analysis. This approach is convenient because unlike state-of-the-art (SoA) models, our proposed model solves different situations that may occur in the video scene (severe dynamic background, initial frames with dynamic objects, static background, stationary objects, etc.) with a specialized module. Furthermore, TMSA_SOM automatically identifies whether the scene has drastically changed (e.g., stationary objects of interest become dynamic or drastic illumination changes have occurred) and automatically detects when the scene has become stable again and uses this information to update the background model in a fast way. The proposed model was validated with three different video databases: Change Detection, BMC, and Wallflower. Findings showed a very competitive performance considering metrics commonly used in the literature to compare SoA models. TMSA_SOM also achieved the best results on two perceptual metrics, Ssim and D-Score, and obtained the best performance on the global quality measure, FSD (based on F-Measure, Ssim, and D-Score), demonstrating its robustness with different and complicated non-controlled scenarios. TMSA_SOM was also compared against SoA neural network approaches obtaining the best average performance on Re, Pr, and F-Measure.

Adaptive Pedestrian Detection in Infrared Images Using Background Subtraction and Local Thresholding

Infrared (IR) imaging is the order of the day with potential real life applications such as surveillance, defence, non-military applications and so on. Low contrast, poor illumination due to capturing devices and moderate to low environmental conditions are the general characterizations of IR images. In addition, the occlusion of objects make the detection more challenging. The objects considered in this paper are pedestrians. A simple and efficient single image handling pedestrian detection method is proposed in this paper. The two major tasks in the proposed method are background subtraction model and local adaptive thresholding. The major contribution of the paper is the adaptive calculation of the required parameters based on the image characteristics. Experiments are conducted on the standard OSU thermal pedestrian database to show the robustness of the proposed method. The proposed method attain detection rate of 90% under various environmental conditions which is superior than the other existing single image handling methods.

Background Subtraction: Theory and Practice

Background subtraction is a widely used concept for detection of moving objects in videos. In the last two decades there has been a lot of development in designing algorithms for background subtraction, as well as wide use of these algorithms in various important applications, such as visual surveillance, sports video analysis, motion capture, etc. Various statistical approaches have been proposed to model scene backgrounds. The concept of background subtraction also has been extended to detect objects from videos captured from moving cameras. This book reviews the concept and practice of background subtraction. We discuss several traditional statistical background subtraction models, including the widely used parametric Gaussian mixture models and non-parametric models. We also discuss the issue of shadow suppression, which is essential for human motion analysis applications. This book discusses approaches and tradeoffs for background maintenance. This book also reviews many of the recent developments in background subtraction paradigm. Recent advances in developing algorithms for background subtraction from moving cameras are described, including motion-compensation-based approaches and motion-segmentation-based approaches. For links to the videos to accompany this book, please see sites.google.com/a/morganclaypool.com/backgroundsubtraction/ Table of Contents: Preface / Acknowledgments / Figure Credits / Object Detection and Segmentation in Videos / Background Subtraction from a Stationary Camera / Background Subtraction from a Moving Camera / Bibliography / Author's Biography

Arbitrary cylinder color model for the codebook based background subtraction.

The codebook background subtraction approach is widely used in computer vision applications. One of its distinguished features is the cylinder color model used to cope with illumination changes. The performances of this approach depends strongly on the color model. However, we have found this color model is valid only if the spectrum components of the light source change in the same proportion. In fact, this is not true in many practical cases. In these cases, the performances of the approach would be degraded significantly. To tackle this problem, we propose an arbitrary cylinder color model with a highly efficient updating strategy. This model uses cylinders whose axes need not going through the origin, so that the cylinder color model is extended to much more general cases. Experimental results show that, with no loss of real-time performance, the proposed model reduces the wrong classification rate of the cylinder color model by more than fifty percent.

Automatic Detection of Operator without Safety Helmet

It is very important to protect the safety of the human head with helmet. Traditional detection for helmet wearing mainly relies on manual approach, which was more subjective that a missing condition may happen caused by fatigue and other factors. Owing to this situation, this paper proposed a method for automatic detection of operator without helmet in real-time. Firstly, Gaussian model for background subtraction is used to detect moving target. Secondly, HOG feature extraction can be used to classify the human target from vehicle. Then, a color feature extraction algorithm is proposed for helmet recognition. The algorithm has been applied into the real time monitoring system and verified with higher accuracy.

Video Background Subtraction in Complex Environments

Background subtraction models based on mixture of Gaussians have been extensively used for detecting objects inmotion in a wide variety of computer vision applications. However, background subtraction modeling is still an openproblem particularly in video scenes with drastic illumination changes and dynamic backgrounds (complexbackgrounds). The purpose of the present work is focused on increasing the robustness of background subtractionmodels to complex environments. For this, we proposed the following enhancements: a) redefine the modeldistribution parameters involved in the detection of moving objects (distribution weight, mean and variance), b)improve pixel classification (background/foreground) and variable update mechanism by a new time-space dependentlearning-rate parameter, and c) replace the pixel-based modeling currently used in the literature by a new space-timeregion-based model that eliminates the noise effect caused by drastic changes in illumination. Our proposed schemecan be implemented on any state of the art background subtraction scheme based on mixture of Gaussians toimprove its resilient to complex backgrounds. Experimental results show excellent noise removal and object motiondetection properties under complex environments.

Adaptation of known background subtraction methods in the case of a moving PTZ camera mounted on a mobile platform

A general method of adaptation of pixel-by-pixel background subtraction models designed for a fixed camera is developed in the case of a PTZ camera mounted on a mobile platform. The method involves the use of two identical pixellevel background models. The first one is directly applied to the classification of the current scene, while the second one is used to prepare the first model for the transformation to the new coordinate system, the one closer to the current position of the PTZ video camera. The chosen solution makes it possible to eliminate a large number of false positives, which inevitably occur after each transformation of the background model. The experimental verification of the developed method using two well-known background models GMM and ViBe demonstrated good quality of the scene classification and low computational load.

MANUAL MODEL BACKGROUND SUBTRACTION METHOD FOR RADAR-BASED BREAST CANCER DETECTION EXPERIMENT

We propose a manual model background subtraction method to detect the breast cancer tumor. A commercially made synthetic breast phantom model is used in our experiment. Ultra-wideband (UWB) antennas have been applied to the simulation experiment of breast cancer detection. The method is effective to detect the breast cancer tumor in real patients in the future.

Segmentation of sows in farrowing pens

The correct segmentation of a foreground object in video recordings is an important task for many surveillance systems. The development of an effective and practical algorithm to segment sows in grayscale video recordings captured under commercial production conditions is described. The segmentation algorithm combines a modified adaptive Gaussian mixture model for background subtraction with the boundaries of foreground objects, which is obtained by using dyadic wavelet transform. This algorithm can accurately extract the shapes of a sow under complex environments, such as dynamic background and illumination changes as well as motionless foreground objects. About 97% of the segmented binary images in the validation data sets can be used to track sow behaviours, such as position, orientation and movement. The experimental results demonstrate that the proposed algorithm is able to provide a basis to detect and classify the sow behaviours in farrowing pens.