Review Paper on Moving Object Tracking In Video Processing

IJSRD - International Journal for Scientific Research & Development| Vol. 3, Issue 01, 2015 | ISSN (online): 2321-0613 Review Paper on Moving Object Tracking in Video Processing Jaydeep S. Dadhaniya1 Prof. Kirit D. Bhalsod2 1 Student of M.E 2Professor 1,2 Department of Electronics & Communication Engineering 1,2 Nobel Group of Institution-Junagadh, Gujarat, India Abstract— Real time moving object detection and tracking is one of the important research fields that have gained a lot of attention in the last few years. Tracking is required for security, safety and site management. Cameras installed around us but there are no means to monitor all of them continuously. It is necessary to develop technologies that automatically process those images in order to detect problematic situations or unusual behavior of human or object. Design computer vision base automated video surveillance system addresses real-time observation of object within a busy environment leading to the description of their actions and interactions. Object detection by background subtraction technique. Using single camera we detect and track human behavior. Background subtraction is the process of separating out the foreground objects from the background in a sequence of video frames. If human entity is cross the line design security in mall or public area the object is tracked. It is laborious to track and trace people over multiple cameras. In this paper, we present review for some system for real-time tracking and fast interactive retrieval of persons in video streams from single static surveillance camera. Key words: Object Tracking, Video Processing I. INTRODUCTION A. Image Processing: Image processing is any form of signal processing for which the input is an image, such as a photograph or video frame; the output of image processing may be either an image or a set of characteristics or parameters related to the image. Image processing is a method to convert an image into digital form and perform some operations on it, in order to get an enhanced image or to extract some useful information from it. It is a type of signal dispensation in which input is image, like video frame or photograph and output may be image or characteristics associated with that image. Usually Image Processing system includes treating images as two dimensional signals while applying already set signal processing methods to them. II. LITERATURE REVIEW Recognition of an object from a scene has been one of the most researched fields. Numerous algorithms have been proposed for object detection but only few of them have acceptable success rate. Further, the real world implementation of these systems is bounded by constraints such as fixed or white background. The primary focus of this paper is to discuss an innovative technique for object detection and object tracking in an different properties of image in area like brightness, luminance. The object image property can vary with a time and containing numerous objects of different shape and sizes. This greatly increases the usability of the application in a real world environment rather than for a laboratory-like environment. This paper deals with enumerating the steps involved in the process of object detection and tracking. Isaac Cohen & Gerard Medioni [1] In this paper they address the problem of detection and tracking of moving objects in a video stream obtained from a moving airborne platform. The proposed method relies on a graph representation of moving objects which allows deriving and maintaining a dynamic template of each moving object by enforcing their temporal coherence. This inferred template along with the graph representation used in our approach allows us to characterize objects trajectories as an optimal path in a graph. The proposed tracker allows dealing with partial occlusions, stopping and going motion in very challenging situations. They demonstrate results on a number of different real sequences. Then they define an evaluation methodology to quantify our results and show how tracking overcome detection errors. Yu Zhong, Anil K. Jain, [2] in this paper they propose a novel method for object tracking using prototypebased deformable template models. To track an object in an image sequence, they use a criterion which combines two terms: the frame-to-frame deviations of the object shape and the fidelity of the modeled shape to the input image. The deformable template model utilizes the prior shape information which is extracted from the previous frames along with a systematic shape deformation scheme to model the object shape in a new frame. The following image information is used in the tracking process: 1) edge and gradient information: the object boundary consists of pixels with large image gradient, 2) region consistency: the same object region possesses consistent color and texture throughout the sequence, and 3) interframe motion: the boundary of a moving object is characterized by large interframe motion. The tracking proceeds by optimizing an objective function which combines both the shape deformation and the fidelity of the modeled shape to the current image (in terms of gradient, texture, and interframe motion). The inherent structure in the deformable template, together with region, motion, and image gradient cues, makes the proposed algorithm relatively insensitive to the adverse effects of weak image features and moderate amounts of occlusion. Karthik Hariharakrishnan and Dan Schonfeld [3] In this paper they propose a fast object tracking algorithm that predicts the object contour using motion vector information. The segmentation step common in region-based tracking methods is avoided, except for the initialization of the object. Tracking is achieved by predicting the object boundary using block motion vectors followed by updating the contour using occlusions/disocclusion detection. An adaptive block-based approach has been used for estimating motion between frames. An efficient modulation scheme is used to control the gap between frames used for motion estimation. The algorithm m for detecting disocclusion proceeds in two steps. First, uncovered regions are estimated All rights reserved by www.ijsrd.com 42 Review Paper on Moving Object Tracking in Video Processing (IJSRD/Vol. 3/Issue 01/2015/013) from the displaced frame difference. These uncovered regions are classified into actual disocclusion and false alarms by observing the motion Characteristics of uncovered regions. Occlusion and disocclusion are considered as dual events and this relationship is explained in detail. The algorithm for detecting occlusion is developed by modifying the disocclusion detection algorithm in accordance with the duality principle. The overall tracking algorithm is computationally superior to existing region-based methods for object tracking. The immediate applications of the proposed tracking algorithm are video compression using MPEG-4 and content retrieval based on standards like H.264. Preliminary simulation results demonstrate the performance of the proposed algorithm. Kaiqi Huang, Liangsheng Wang, Tieniu Tan & Steve Maybank [4] In this paper, Autonomous video surveillance and monitoring has a rich history. Many deployed systems are able to reliably track human motion in indoor and controlled outdoor environments. However, object detection and tracking at night remain very important problems for visual surveillance. The objects are often distant, small and their signatures have low contrast against the background. Traditional methods based on the analysis of the difference between successive frames and a background frame will do not work. In this paper, a novel real time object detection algorithm is proposed for nighttime visual surveillance. The algorithm is based on contrast analysis. In the first stage, the contrast in local change over time is used to detect potential moving objects. Then motion prediction and spatial nearest neighbour data association are used to suppress false alarms. Experiments on real scenes show that the algorithm is effective for night-time object detection and tracking. Gustavo Moreira, Bruno Feijo, Helio Lopes & Raul Queiroz Feitosa [5] in this paper the author propose an integrated detection and tracking method suitable for highdefinition videos at real-time frame rates. In this method they implement a frame segmentation procedure using integral images of the background, which permits to discard the analysis of several image parts of each frame and achieve high frame rates. Also they extend the proposed algorithm to detect multiple objects in parallel. Shashank Prasad, Shubhra Sinha [6] According to this paper, In recent years, there has been extensive research in the field of object detection and tracking. Many remarkable algorithms have been developed for object detection and tracking, including color segmentation, edge tracking and many more. However, all these algorithms faced the limited success in their implementation in the real world and were also bounded by the constraints such as white/plain background. This paper is the result of our research where our research team developed and implemented object detection and tracking system operational in an unknown background, using real-time video processing and a single camera. The proposed system has been extensively tested to operate in complex, real world, non-plain, light variant, changing background. Giannis Chantas, Nikos Nikolaidis and Ioannis Pitas [7] In this paper a general Bayesian post-processing methodology for performance improvement of object tracking in stereo video sequences is proposed. They utilize the results of any single channel visual object tracker in a Bayesian framework, in order to refine the tracking accuracy in both stereo video channels. In this framework, a variational Bayesian algorithm is employed, where prior knowledge about the object displacement (movement) is incorporated via a prior distribution. This displacement information is obtained in a pre-processing step, where object displacement is estimated via feature extraction and matching. In parallel, disparity information is extracted and utilized in the same framework. The improvements introduced by the proposed methodology in terms of tracking accuracy are quantified through experimental analysis. Mohammad Alfraheed [8] According to this paper an autonomous driving, object tracking is necessary to gather actual information about the object of interest. The longitudinal and lateral controls of automated highway systems need a target object not only to maintain the safety distance between vehicles but also to keep the following vehicle in the same track as the preceding vehicle. So far automated highway systems were only developed for urban and highway environment depending on lane markings. In future, their application should be extended to unstructured environments (e.g. desert) and be adapted for heterogeneous vehicles. In this paper an approach towards this is presented, where the back view of preceding vehicle is the target object. This solution is independent from the environmental structure as well as additional equipment like infrared emitters. In this paper, the tracking process of the back view is discussed using video streams recorded by a stereo vision system. For an accurate and fast tracking in unstructured environment and with heterogeneous platoons the proposed method is a supplement to the detection process. Therefore, the tracking process has to be a) applicable under real time constraints and b) adaptable in dynamic environments. Compared to other methods related to object detection and tracking, the proposed method reduces the running time for the tracking of the back view from reported 12 - 30 to 16 66 frame/s. III. CONCLUSION OF ENTIRE LITERATURE REVIEW After studying above research paper and thesis following conclusions derived 1) There are several problems related to the analysis of a video stream. The framework proposed is based on a graph representation of the moving regions extracted from a video acquired by a moving platform. The integration of the detection and tracking in this graph representation allows to dynamically inferring a template of all moving objects in order to derive a robust tracking in situations such as stop and go motion and partial occlusion. Finally, the quantification of the results through the definition of the metrics DR and FAR provides a confidence measure characterizing the reliability of each extracted trajectory. The obtained results will be improved by further processing the false alarms in order to discard the trajectories due to regions with good temporal coherence which does not correspond to moving objects, and these are, typically, regions due to strong parallax. 2) The proposed framework is quite general and can be applied to a number of tracking tasks. Future work will All rights reserved by www.ijsrd.com 43 Review Paper on Moving Object Tracking in Video Processing (IJSRD/Vol. 3/Issue 01/2015/013) incorporate temporal prediction such as Kalman filtering to improve the tracking results. They will also investigate how to weigh the different image cues based on the uncertainty in estimating their values. 3) In this paper, they have proposed a simple tracking algorithm that avoids segmentation except for initialization of the object partition during the initial frame. Object tracking using block motion vectors have seldom been exploited. Such an approach can be implemented using parallel processors and hence lends itself to a real-time implementation. The goal was to develop an algorithm that extracts video objects whose accuracy is close to the region-based approaches reported in the literature and at the same time perform tracking in a computationally efficient manner. Occlusion and disocclusion are viewed as reverse problems. An efficient algorithm for detecting occlusions is proposed and modified in accordance with the duality principle to develop a disocclusion algorithm. As the object mask is modified to take care of occlusions/disocclusion, the object can be tracked accurately for a longer time without requiring reinitialization/ re-segmentation. The tracking algorithm proposed inherently can be extended in a straightforward manner to extract multiple objects. The tracking algorithm, however, performs poorly for objects that are relatively small and relevant changes are currently being investigated. The tracking obtained using the method can be used as a predicted position for methods that employ snakes to track contours. The prediction can be based on the computed affine model and the innovations process can be used to correct for errors in the prediction. This approach is similar to the Kalman filter based approaches. This approach is also being considered for future research. Motion estimation using the sum of absolute differences does not deal with Gaussian noise and hence the motion vectors tend to be erroneous causing the tracked result to deteriorate. This is another aspect that requires study. 4) Object detection and tracking at night is very important for night surveillance, which is key part of 24 h visual surveillance. In this paper, they proposed object detection and tracking algorithm for night surveillance based on inter-frame differences. Object detection is based on local contrast changes and detection results are improved by tracking the detected objects from one frame to the next. Experiments demonstrate that our algorithm has the ability to detect and track objects robustly at night under conditions in which more conventional algorithms fail. There are several parameters and thresholds in the new algorithm. Some parameters are adjusted adaptively, for example, the threshold to determine significant inter-frame differences and the threshold on the differences between contrast scores. Other parameters such as the size of rectangular region for contrast measure, the threshold on contrast measure and the threshold on the distance measure between two rectangles are chosen by hand. In the future, we will use a multi-scale algorithm, similar to those used in face detection, to decide the size of the rectangular 5) 6) 7) 8) region for contrast measure. The threshold on contrast measure will be decided by a learning algorithm. These methods for computing thresholds automatically will improve our system greatly. In this paper they extend Viola and Jones‟ detection algorithm towards a real-time integrated detection and kernel tracking algorithm. Firstly they introduce the idea of using the integral image of the background to discard from the analysis several parts of each frame. Furthermore, this discarding process reveals an adaptive frame segmentation that defines a reduced area for object detection. Another contribution is to expand these ideas to deal with multiple objects in parallel. Finally, our method presents high performance in terms of processing time without missing the qualities of Viola and Jones‟ original algorithm. The proposed algorithm for object detection and tracking in unknown environment was extensively tested to operate in complex, real world; non-plain and changing background was found to possess remarkable accuracy and precision of 99%. Research team has tested the proposed algorithm to track assorted objects against an environment consisting of cluttered objects of varying sizes, shapes and colors. The implementation of the algorithm was found be extremely fast and robust. It also made tracking of objects highly feasible in light variant conditions. The proposed algorithm is the first algorithm that addressed seamless tracking of object of color „c‟ against an identical background of color „c‟. Its accuracy in tracking objects camouflaged in the background was found to be 99.9%. Thus, the proposed algorithm for object detection and tracking in unknown environment shall open new vista in field of computer vision for developing real world applications and also improvising currently existing algorithms to be operational in the real world. An object tracking Bayesian post-processing methodology for stereoscopic sequences was presented in this paper. The methodology refines the outputs of standard tracking algorithms, by exploiting, the left and right channel tracking results. Moreover, object displacement over time, as well as disparity information, were exploited successfully to this end. The refined tracking results are significantly better than those provided by the initial, single-channel tracking algorithm. In the future, we plan to improve the stochastic model. The proposed method has been tested with over 1430 frames. Based on the distance of the preceding vehicle the results have been clustered. In context of reliable property (running under real time constraints and environmental effects), the successful results achieved for a distance less than 10 meters were 100 %. For distance between 10 and 12 meters, the successful results are 91 %. Regarding the safety distance, 91 % is enough to warn the core system of the platoon in order to decrease the distance to less than or equal 10 meters. Problems occurred for a distance larger than 12 meters because the back view is not clear enough to be detected. However, the successful results of the last All rights reserved by www.ijsrd.com 44 Review Paper on Moving Object Tracking in Video Processing (IJSRD/Vol. 3/Issue 01/2015/013) cluster (arithmetic average 61%) could be improved using the detection process whenever the tracking process did not track the BVPV. The proposed method is distinguished by its ability to rapidly track the BVPV. Moreover, it allows working under real time constraints, because the running time lies around 16 66 frames/s. REFERENCES [1] Isaac Cohen, Gerard Medioni, Detecting and Tracking Moving Objects for Video Surveillance, University of Southern California Institute for Robotics and Intelligent Systems, IEEE Proc. Computer Vision and Pattern Recognition, Jun. 2325, 1999 [2] Yu Zhong, Anil K. Jain and M.-P. Dubuisson-Jolly, Object Tracking Using Deformable Templates, IEEE trans. on Pattern Analysis and Machine Intelligence, VOL. 22, NO. 5, May 2000 [3] Karthik Hariharakrishnan and Dan Schonfeld, Fast Object Tracking Using Adaptive Block Matching, IEEE Transaction on Multimedia, VOL. 7, NO. 5, October 2005 [4] Kaiqi Huang, Liangsheng Wang, Tieniu Tan, Steve Maybank, A real-time object detecting and tracking system for outdoor night surveillance, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100080, China, School of Computer Science and Information Systems, Birkbeck College, Malet Street, London WC1E 7HX, UK, 23 May 2007 [5] Gustavo Moreira, Bruno Feijo, H´elio Lopes, Raul Queiroz Feitosa, Real-time Object Tracking in High-Definition Video Using Frame Segmentation and Background Integral Images, Departamento de Informatica, Departamento de Engenharia El‟etrica Pontif´ıcia Universidade Cat´olica do Rio de Janeiro, XXVI Conference on Graphics, Patterns and Images,2013 [6] Shashank Prasad, Shubhra Sinha, Real-time Object Detection and Tracking in an Unknown Environment, Mumbai University, World Congress on Information and Communication Technologies, page no.-1056-1061, 2011 [7] Giannis Chantas, Nikos Nikolaidis and Ioannis Pitas, A Bayesian Methodology For Visual Object Tracking on Stereo Sequences, Department of Informatics, Aristotle University of Thessaloniki, Box 451, Thessaloniki, GR 54124, Greece, 2013 [8] Mohammad Alfraheed, “Automated Heterogeneous Platoons in Unstructured Environment: Real Time Tracking of a Preceding Vehicle Using Video Stream”, Department of Mathematics and Computer Science, Tafila Technical University, Tafila – Jordan, IEEE, 2014 All rights reserved by www.ijsrd.com 45