Showing papers on "Change detection published in 2000"

Automatic analysis of the difference image for unsupervised change detection

Abstract: One of the main problems related to unsupervised change detection methods based on the "difference image" lies in the lack of efficient automatic techniques for discriminating between changed and unchanged pixels in the difference image. Such discrimination is usually performed by using empirical strategies or manual trial-and-error procedures, which affect both the accuracy and the reliability of the change-detection process. To overcome such drawbacks, in this paper, the authors propose two automatic techniques (based on the Bayes theory) for the analysis of the difference image. One allows an automatic selection of the decision threshold that minimizes the overall change detection error probability under the assumption that pixels in the difference image are independent of one another. The other analyzes the difference image by considering the spatial-contextual information included in the neighborhood of each pixel. In particular, an approach based on Markov Random Fields (MRFs) that exploits interpixel class dependency contexts is presented. Both proposed techniques require the knowledge of the statistical distributions of the changed and unchanged pixels in the difference image. To perform an unsupervised estimation of the statistical terms that characterize these distributions, they propose an iterative method based on the Expectation-Maximization (EM) algorithm. Experimental results confirm the effectiveness of both proposed techniques.

Adaptive filtering and change detection

Abstract: INTRODUCTION Extended Summary. Applications. SIGNAL ESTIMATION On--Line Approaches. Off--Line Approaches. PARAMETER ESTIMATION Adaptive Filtering. Change Detection Based on Sliding Windows Change Detection Based on Filter Banks STATE ESTIMATION Kalman Filtering Change Detection Based on Likelihood Ratios Change Detection Based on Multiple Models Change Detection Based on Algebraical Consistency Tests THEORY Evaluation Theory Linear Estimation A. Signal models and notation B. Fault detection terminology

Performance characterization of video-shot-change detection methods

Abstract: A number of automated shot-change detection methods for indexing a video sequence to facilitate browsing and retrieval have been proposed. Many of these methods use color histograms or features computed from block motion or compression parameters to compute frame differences. It is important to evaluate and characterize their performance so as to deliver a single set of algorithms that may be used by other researchers for indexing video databases. We present the results of a performance evaluation and characterization of a number of shot-change detection methods that use color histograms, block motion matching, or MPEG compressed data.

Relative Radiometric Normalization Performance for Change Detection from Multi-Date Satellite Images

Abstract: Relative radiometric normalization (RRN minimizes radiometric differences among images caused by inconsistencies of acquisition conditions rather than changes in sudace reflectance. Five methods of RRN have been applied to 1973, 1983, and 1988 Landsat MSS images of the Atlanta area for evaluating their pedormance in relation to change detection. These methods include pseudoinvariant features (PIF), radiometric control set (RCS), image regression (ml, no-change set determined from scattergrams (NC], and histogram matching (MM), all requiring the use of a reference-subject image pair. They were compared in terms of their capability to improve visual image quality and statistical robustness. The way in which different RRN methods affect the results of information extraction in change detection was explored. It was found that RRN methods which employed a large sample size to relate targets of subject images to the reference image exhibited a better overall performance, but tended to reduce the dynamic range and coefficient of variation of the images, thus undermining the accuracy of image classification. It was also found that visually and statistically robust RRN methods tended to substantially reduce the magnitude of spectral differences which can be linked to meaningful changes in landscapes. Finally, factors affecting the pedormance of relative radiometric normalization were identified, which include land-use/ land-cover distribution, water-land proportion, topographic relief, similarity between the subject and reference images, and sample size.

Wetland change detection on the Kafue Flats, Zambia, by classification of a multitemporal remote sensing image dataset

Abstract: The paper describes a remote sensing change detection approach used to assess change on a section of the Kafue Flats floodplain wetland system in southern Zambia, which is under the pressures of reduced regional rainfall and damming and water abstraction by man. Four images from September 1984 (Landsat MSS), 1988 (Landsat MSS), 1991 (Landsat TM) and 1994 (Landsat TM) were used. Being near-anniversary images, the change detection error introduced by mere seasonal differences was minimized. Following atmospheric correction of the reference (1994) image, the images were radiometrically normalized and geometrically registered to a common map projection. Each image was separately classified into categories of open water, dense green vegetation, sparse green vegetation, very sparse green vegetation, dry and burnt land. Similar, supervised maximum likelihood classification procedures were employed on all images. The classified images produced were analysed for change in each land-cover category by overlaying the...

Visual Search for Change: A Probe into the Nature of Attentional Processing

Abstract: A set of visual search experiments tested the proposal that focused attention is needed to detect change. Displays were arrays of rectangles, with the target being the item that continually changed its orientation or contrast polarity. Five aspects of performance were examined: linearity of response, processing time, capacity, selectivity, and memory trace. Detection of change was found to be a self-terminating process requiring a time that increased linearly with the number of items in the display. Capacity for orientation was found to be about five items, a value comparable to estimates of attentional capacity. Observers were able to filter out both static and dynamic variations in irrelevant properties. Analysis also indicated a memory for previously attended locations. These results support the hypothesis that the process needed to detect change is much the same as the attentional process needed to detect complex static patterns. Interestingly, the features of orientation and polarity were found to be handled in somewhat different ways. Taken together, these results not only provide evidence that focused attention is needed to see change, but also show that change detection itself can provide new insights into the nature of attentional processing.

Semantic Informativeness Mediates the Detection of Changes in Natural Scenes

Abstract: Three experiments investigated whether the semantic informativeness of a scene region (object) influences its representation between successive views. In Experiment 1, a scene and a modified version of that scene were presented in alternation, separated by a brief retention interval. A changed object was either semantically consistent with the scene (non-informative) or inconsistent (informative). Change detection latency was shorter in the semantically inconsistent versus consistent condition. In Experiment 2, eye movements were eliminated by presenting a single cycle of the change sequence. Detection accuracy was higher for inconsistent versus consistent objects. This inconsistent object advantage was obtained when the potential strategy of selectively encoding inconsistent objects was no longer advantageous (Experiment 3). These results indicate that the semantic properties of an object influence whether the representation of that object is maintained between views of a scene, and this influence is not caused solely by the differential allocation of eye fixations to the changing region. The potential cognitive mechanisms supporting this effect are discussed.

The role of iconic memory in change detection tasks

Abstract: In three experiments, subjects attempted to detect the change of a single item in a visually presented array of items. Subjects' ability to detect a change was greatly reduced if a blank interstimulus interval (ISI) was inserted between the original array and an array in which one item had changed ('change blindness'). However, change detection improved when the location of the change was cued during the blank ISI. This suggests that people represent more information of a scene than change blindness might suggest. We test two possible hypotheses why, in the absence of a cue, this representation fails to produce good change detection. The first claims that the intervening events employed to create change blindness result in multiple neural transients which co-occur with the to- be-detected change. Poor detection rates occur because a serial search of all the transient locations is required to detect the change, during which time the representation of the original scene fades. The second claims that the occurrence of the second frame overwrites the representation of the first frame, unless that information is insulated against overwriting by attention. The results support the second hypothesis. We conclude that people may have a fairly rich visual representation of a scene while the scene is present, but fail to detect changes because they lack the ability to simultaneously represent two complete visual representations.

Fast scene change detection using direct feature extraction from MPEG compressed videos

Abstract: In order to process video data efficiently, a video segmentation technique through scene change detection must be required. This is a fundamental operation used in many digital video applications such as digital libraries, video on demand (VOD), etc. Many of these advanced video applications require manipulations of compressed video signals. So, the scene change detection process is achieved by analyzing the video directly in the compressed domain, thereby avoiding the overhead of decompressing video into individual frames in the pixel domain. In this paper, we propose a fast scene change detection algorithm using direct feature extraction from MPEG compressed videos, and evaluate this technique using sample video data, First, we derive binary edge maps from the AC coefficients in blocks which were discrete cosine transformed. Second, we measure edge orientation, strength and offset using correlation between the AC coefficients in the derived binary edge maps. Finally, we match two consecutive frames using these two features (edge orientation and strength). This process was made possible by a new mathematical formulation for deriving the edge information directly from the discrete cosine transform (DCT) coefficients. We have shown that the proposed algorithm is faster or more accurate than the previously known scene change detection algorithms.

New enhancements to cut, fade, and dissolve detection processes in video segmentation

Abstract: We present improved algorithms for cut, fade, and dissolve detection which are fundamental steps in digital video analysis. In particular, we propose a new adaptive threshold determination method that is shown to reduce artifacts created by noise and motion in scene cut detection. We also describe new two-step algorithms for fade and dissolve detection, and introduce a method for eliminating false positives from a list of detected candidate transitions. In our detailed study of these gradual shot transitions, our objective has been to accurately classify the type of transitions (fade-in, fade-out, and dissolve) and to precisely locate the boundary of the transitions. This distinguishes our work from other early work in scene change detection which tends to focus primarily on identifying the existence of a transition rather than its precise temporal extent. We evaluate our improved algorithms against two other commonly used shot detection techniques on a comprehensive data set, and demonstrate the improved performance due to our enhancements.

Adaptive change detection for real-time surveillance applications

Abstract: This paper describes a new real-time approach for detecting changes in grey level image sequences, which were taken from stationary cameras. This new method combines a temporal difference method with an adaptive background model subtraction scheme. When changes in illumination occur the background model is automatically adapted to suit the new conditions. For the adaptation of the background model a new method is proposed, which avoids reinforcement of adaptation errors by performing the adaptation solely on those regions that were detected by the temporal difference method rather than using the regions resulting from the overall algorithm. Thus the adaptation process is separated from the results of its own background subtraction algorithm. The change detector was successfully tested both in a vision-based workspace monitoring system for different kinds of non-autonomous service robots and in a surveillance scenario, in which it was the task to detect people in a subway-platform scenario. The proposed real-time algorithm showed recognition rates of up to 90% in the foreground and 84% in the background and performed in all cases at least 12% better than the alternative method of adaptive background estimation which uses a modified Kalman filtering technique.

The impact of misregistration upon composited wide field of view satellite data and implications for change detection

Abstract: Composited wide field of view satellite data are used for many applications and increasingly for studies of global change. Although several compositing schemes have been suggested, all assume perfect geometric registration, which is not operationally feasible. In this study, models of the satellite imaging, geometric registration, and compositing processes are used to investigate the impact of misregistration upon the position of high contrast edges found in composited satellite data. Simulations are performed with respect to the compositing of advanced very high resolution radiometer (AVHRR) and moderate resolution imaging spectroradiometer (MODIS) data. Contrast edges are found to be systematically shifted in maximum and minimum value composites. The degree of shifting increases with the number of orbits that are composited, the degree of misregistration and the view zenith angle. The implications of these findings upon the utility of composited satellite data for change detection are discussed. The shifts may systematically bias estimates of location and area when composited data are used. They may also cause small and/or fragmented features, which are evident in individual orbits to disappear in composited data, precluding the ability to map such features or to detect their occurrence under a change detection scheme.

Illumination-invariant change detection

Abstract: Moving objects in image sequences acquired by a static camera can be detected by analyzing the grey-level difference between successive frames. Direct motion detection, however, will also detect fast variations of scene illumination. This paper describes a method for motion detection that is considerably less sensitive to time-varying illumination. It is based on combining a motion detection algorithm with an homomorphic filter which effectively suppresses variable scene illumination. To this end, the acquired image sequence is modelled as being generated by an illumination and a reflectance component that are approximately separated by the filter. Detection of changes in the reflectance component is directly related to scene changes, i.e., object motion. Real video data are used to illustrate the system's performance.

Comparison of SAR amplitude vs. coherence flood detection methods - a GIS application

Abstract: Flood area detection from multipass Synthetic Aperture Radar (SAR) data can be performed via amplitude change detection techniques. These methods allow flooded zones to be discriminated only when they are flooded at the time of the second passage, and not at the time of the first one. Coherence derived from multipass SAR interferometry can be used instead, as an indicator of changes in the electromagnetic scattering behaviour of the surface, thus potentially revealing all the areas affected by the flood event at any time between the two passes. The paper presents a prototype application of such techniques, that is, a flood map obtained from ERS-1/2 data taken over Beziers (Southern France), through proper thresholding of a combination of amplitude and coherence information. Produced in the framework of an ESA project, the map consists of a DXF vector file which can be imported directly into most commercial GIS software.

Detecting Changes in Novel, Complex Three-dimensional Objects

Abstract: Four experiments assessed change detection performance for displays consisting of a single, novel, multipart object, leading to several new findings. First, larger changes (involving more object parts) were more difficult to detect than smaller changes. Second, change detection performance for displays of a temporarily occluded moving object was no more or less sensitive than detection performance for displays of static objects disappearing and reappearing; however, item analyses did indicate that detection may have been based on different representations in these two situations. Third, training observers to recognize objects before the detection task had no measurable effect on sensitivity levels, but induced different biases depending on the training conditions. Finally, some participants' performance revealed implicit change detection on trials in which they explicitly responded that they saw no change.

Accuracy Assessment Curves for Satellite-Based Change Detection

Abstract: column (Lillesand and Kiefer, 1994, p. 613; Story and ConThe concept of "accuracy assessment curves" as they relate to galton, 1986; Aronoff, 1982a; Aronoff, 1982b)), user's accuracy image-based change detection is developed. These curves are (the classified for a given row a graphical representation of the relationship between typical divided by the total for that row "users accuracy", (Lillesand accuracy assessment figures and a binary change detection and Kiefer119949 PP. 613; Story and Congalton* 1986; Aronoff9 threshold level. After an introduction to accuracy assessment 1982a; o Congalton the binary change threshold level is first shown with a series et ~l.9 1983)). shows a standard matrix. of error matrices. This relationship is then expanded across a For binary change detection, the image processing range of thresholds and shown as accuracy assessment curves. is a metric tobeusedto quantify changesin These curves show how accuracy figures are dependent on the reflectance values between two images- This is typically some change threshold level. hi^ dependency indicates distance function between corrected reflectance values or some limitations of binary &ange maps. These limitations between index values derived from each image. For any binary provide the impetus for a continuous change product. We change detection* the accuracy figures will change depending describe how a continuous Probability-of-change (PW) image, the level (Fung and LeDrew, 1988). We try to derived frorn statistical models, can be incorporated with the express this dependence in an intuitive way with the followaccuracy assessment curves to produce a more meaningful and ing informative change detection product.

Applying observer based FDI techniques to detect faults in dynamic and bounded stochastic distributions

Abstract: This paper presents novel approach on applying the standard observer based fault detection technique to the change detection of the output probability density functions for dynamic stochastic syste...

Video object extraction based on adaptive background and statistical change detection

Abstract: This paper introduces a system for video object extraction useful for general applications where foreground objects move within a slow changing background. Surveillance of indoor and outdoor sequences is a typical example. The originality of the approach resides in two related components. First, the statistical change detection used in the system does not require any sophisticated parametric tuning as it is based on a probabilistic method. Second, the change is detected between a current instance of the scene and a reference that is updated continuously to take into account slow variation of the background. Simulation results show that the proposed scheme performs well in extracting video objects, with stability and good accuracy, while being of relative reduced complexity.

Uterine EMG analysis: a dynamic approach for change detection and classification

Abstract: Toward the goal of detecting preterm birth by characterizing events in the uterine electromyogram (EMG), the authors propose a method of detection and classification of events in this signal. Uterine EMG is considered as a nonstationary signal and the authors' approach consists of assuming piecewise stationarity and using a dynamic change detector with no a priori knowledge of the parameters of the hypotheses on the process state to be detected. The detection approach is based on the dynamic cumulative sum (DCS) of the local generalized likelihood ratios associated with a multiscale decomposition using wavelet transform. This combination of DCS and multiscale decomposition was shown to be very efficient for detection of both frequency and energy changes. An unsupervised classification based on the comparison between variance-covariance matrices computed from selected scales of the decomposition was implemented after detection. Finally a class labeling based on neural networks was developed. This algorithm of detection-classification-labeling gives satisfactory results on uterine EMG: in most cases more than 80% of the events are correctly detected and classified whatever the term of gestation.

Toward specification-driven change detection

Abstract: A change-detection methodology based on explicit user requirements in terms of example imagery and false alarm and misclassification probabilities is discussed and applied. A distance measure between texture features is defined, and its ability is illustrated to measure changes in urban areas in high resolution, panchromatic, spaceborne images.

An adaptive parcel-based technique for unsupervised change detection

Abstract: A novel unsupervised technique for the detection of changes in multi-temporal remote sensing images is presented. It adaptively exploits the spatial-contextual information contained in the neighbourhood of each pixel to reduce the effects of noise and hence to increase change-detection accuracy. In addition, the proposed definition of an adaptive pixel neighbourhood allows a precise location of the borders of changed areas.

A simple recursive algorithm for diagnosis of abrupt changes in random signals

Abstract: We address the problem of detecting and isolating abrupt changes in random signals. An asymptotic optimal solution to this problem, which has been proposed in previous works, involve the number of computations at time t which grows to infinity with t. We propose another more realistic criterion, establish a new simple recursive change detection/isolation algorithm, and investigate its statistical properties.

A minimum-cost thresholding technique for unsupervised change detection

Abstract: We propose an automatic thresholding technique for difference images in unsupervised change detection. Such a technique takes into account the different costs that may be associated with commission and omission errors in the selection of the decision threshold. This allows the generation of maps in which the overall change-detection cost is minimized, i.e. the more critical kind of error is reduced according to end-user requirements.

Motion-based video representation for scene change detection

Abstract: We present a scheme for automatically partitioning videos into scenes. A scene is generally referred to as a group of shots taken at the same site. We first propose a motion annotation algorithm based on the analysis of spatiotemporal image volumes. The algorithm characterizes the motions within shots by extracting and analyzing the motion trajectories encoded in the temporal slices of image volumes. A motion-based keyframe computing and selection strategy is thus proposed to compactly represent the content of shots. With these techniques, we further present a scene change detection algorithm by measuring the similarity of the representative keyframes in shots.

Using change-point detection to support artificial neural networks for interest rates forecasting

Abstract: Interest rates are one of the most closely watched variables in the economy. They have been studied by a number of researchers as they strongly affect other economic and financial parameters. Contrary to other chaotic financial data, the movement of interest rates has a series of change points owing to the monetary policy of the US government. The basic concept of this proposed model is to obtain intervals divided by change points, to identify them as change-point groups, and to use them in interest rates forecasting. The proposed model consists of three stages. The first stage is to detect successive change points in the interest rates dataset. The second stage is to forecast the change-point group with the backpropagation neural network (BPN). The final stage is to forecast the output with BPN. This study then examines the predictability of the integrated neural network model for interest rates forecasting using change-point detection.

Morphological Change Detection Algorithms for Surveillance Applications

Abstract: Vision-based systems for remote surveillance usually involve change detection algorithms for intruders, obstacles or irregularities detection In particular, there is a potentially very cost-effective approach to perform inspection with autonomous robot navigation, computer vision, and change detection based on automatic image registration and subtraction In these cases, a model of the working environment is compared with data acquired during the system functioning in order to extract region of interests Real time applications require simple, fast and reliable algorithms and methodologies presented in the literature show that morphological change detection satisfies these requirements In this paper novel morphological algorithms for scene change detection are introduced; the proposed methods allow to obtain a system whose performances are relatively stationary also with varying environmental condition

Detection and identification of change in naturalistic scenes

Abstract: Research using change detection paradigms has demonstrated that only limited scene information remains available for conscious report following initial inspection of a scene Previous researchers have found higher change identification rates for deletions of parts of objects in line drawings of scenes than additions Other researchers, however, have found an asymmetry in the opposite direction for addition/deletion of whole objects in line drawings of scenes Experiment 1 investigated subjects' accuracy in detecting and identifying changes made to successive views of high quality photographs of naturalistic scenes that involved the addition and deletion of objects, colour changes to objects, and changes to the spatial location of objects Identification accuracy for deletions from scenes was highest, with lower identification rates for object additions and colour changes, and the lowest rates for identification of location changes Data further suggested that change identification rates for the presence/a

Audio scene segmentation using multiple features, models and time scales

Abstract: We present an algorithm for audio scene segmentation. An audio scene is a semantically consistent sound segment that is characterized by a few dominant sources of sound. A scene change occurs when a majority of the sources present in the data change. Our segmentation framework has three parts: a definition of an audio scene; multiple feature models that characterize the dominant sources; and a simple, causal listener model, which mimics human audition using multiple time-scales. We define a correlation function that determines correlation with past data to determine segmentation boundaries. The algorithm was tested on a difficult data set, a 1 hour audio segment of a film, with impressive results. It achieves an audio scene change detection accuracy of 97%.

A unified approach to scene change detection in uncompressed and compressed video

Abstract: There is an increasing need to extract key information automatically from video for the purposes of indexing, fast retrieval and scene analysis. To support this vision, reliable scene change detection algorithms must be developed. This paper describes a unified approach for scene change detection in uncompressed and MPEG-2 compressed video statistical properties of each image. An efficient algorithm is proposed to estimate the statistical features in compressed video without full frame decompression and used these features with the uncompressed domain algorithms to identify scene changes in compressed video. Proposed scheme aims at detecting abrupt transitions and gradual transitions in both uncompressed and MPEG-2 compressed video using a single framework. Results on video of various content types are reported and validated. Furthermore, results show that for uncompressed video the accuracy of the detected transition region is above 98% and above 95% for MPEG-2 compressed video.