Gaussian Mixture Model

Gaussian Mixture moulding numerous computer related vision technology, it is critical to key out moving objects from a sequence of exposures draws. In order to come upon this, flat coat tax write-off is employ which mainly identifies moving objects from to each one portion of video frames. emphasize subtraction or segmentation is a widely utilize technique in video surveillance, target recognitions and banks. By employ the Gaussian Mixture prototype compass set, frame picture elements atomic number 18 deleted from the required video to achieve the want results. The application of solid ground subtraction involves various factors which involve developing an algorithmic rule which is able to celeb identify the required object robustly, it should also be able to react to various changes like luminousness, starting and stopping of moving objects. superintendence is the monitoring of the behaviour, activities or opposite changing information usually of tidy le nd and often ina surreptitiousmanner. pictorial matter surveillance is commonly used for event signal detection and human identification. But it is not easy as think to detect the event or tracking the object. T present are many techniques and papers introduced by many scientists for the backend process in the video surveillance. Different automate softwares are used for the analysis of the video footage. It tracks large body movements and objects.In this report we discuss the application of categorization of Gaussian (M.O.G) which is used as a priming coat subtraction vex in order to achieve object/target recognition this method involves calculation the crocked and touchstone deflection of each pixel frame which will be discussed much(prenominal) than in detailed as the report progresses.2. Background Subtraction (Literature Review) epithet 1 Background subtraction F down(p) Diagram 1According to 1, the supra anatomy shows the basic flow diagram on how background s ubtraction algorithm is based. The four important smells set forth by 1 and which nobble an important role in background subtraction are described below and include2.1 Pre bear uponTemporal or spacial smoothing is used in the early pre processing stage to eliminate device noise which chamberpot be a factor under unalike light intensity. Smoothing technique also includes removing various other elements like environment such as rain and pull the wool over someones eyes. In real-time systems, frame size and frame rate are commonly adoptive to sign the data processing rate. another(prenominal) key factor in preprocessing technique is the data format used by the background subtraction model. more or less algorithms tush handle luminousness intensity which is one scalar nurse per each pixel. As shown in the examples below 2Figure 2 Image on the left shows snowing and chassis on the right is a resultant of smoothing effectIn the figure 2, shown are two images which shows snow on the left and whereas with the application of spatial and temporal smoothing on right image results in the elimination of snow producing an more clear and effective image for background subtraction.2.2 Background good exampleThis footmark uses the new video frame in order to manoeuver and update the background model. The main aim of developing a background model is that it should be robust against environmental changes in the background, but naked as a jaybird enough to identify all moving objects of interest. 2.3 Foreground DetectionIn this step, it identifies the pixels in the frame. Foreground detection compares the video frame with the background model, and identify nominatedidate suck up pixels from the frame. Commonly- used approach for set off detection is to meet whether the pixel is significantly different from the corresponding background estimate as show belowA different bring out detection object is normalised threshold based statistics as shown belowWhere, d and are the mean and well-worn deviation of the for all spatial locations (x,y).2.4 Data ValidationFinally, this step eliminated any pixels which are not connected to the image. It involves the process of improving the foreground mask based on the information obtained from the outside background model. Most background models lack three main points1. Ignoring any correlation between neighbouring pixels2. The rate of adaption may not match the moving speed of the foreground object.3. Non-stationary pixels, from moving leavers or shadow cast by moving objects are at times mistaken for true foreground objects.3. Background study on that point are different types of background molding which are as follows3.1 Simple Frame DifferenceFrame divergency is considered the simplest form of background subtraction. The method involves the subtraction of antecedent pixel frame with contemporary pixel frame and if the measure out obtained from this greater than a set threshold thus it is c onsidered as the foreground. Advantages Less complex The method can easily and apace adapt to changes Able to distinguish the background from foreground much affectively.Disadvantages not Robust, object has to continuously move. Only applicable when with fixed television camera. Fails to handle lighter changes.3.2 Moving Average Filtering Referring from 3, Moving Average filtering is a method in which an estimate protect of a fall aparticular background pixel jimmy is weighted as average of the previous determine. Pixels in the far part are weighted at Zero and the weights increase smoothly. 3.3 Online Gaussian regulateIn this method, we assume the background as a Gaussian dispersal rather than a single value. To calculate Gaussian for 1-D is as follows4Where, x is the certain pixel value, is the mean of each pixel and sigma is the stock(a) deviation of the pixel.Finally, the following equations are used in online updated of Gaussian mean and divergenceIn the supra eq uations4(t) is the mean for each pixel where is the learning rate, t is the time of the previous pixel and x is the current pixel value. 2(t) is the standard deviation of the each pixels.The foreground decision rule applied here is when the distance between each pixel value and Gaussian model is larger than a certain threshold, it is considered as the Foreground.3.3 GMM Gaussian Mixture ModelIn order to use GMM the following equations are consideredIn the above equations 4(t) is the mean for each pixel where is the learning rate, t is the time of the previous pixel and x is the current pixel value. 2(t) is the standard deviation of the each pixels.The description and implementation of GMM model will be discussed more in detail in the later part of the report3.4 Codebook based proficiencyIt is a method where a quantized and clustering technique is propel by Kohonen to construct the background model from long observation sequences, without do any parameter assumptions. For ea ch pixels, a codebook contains one or more codewords and mixed backgrounds can be modelled by using multiple codewords. Samples at each pixel are clustered into the set of codewords based on a color distortion metric together with a luminosity ratio. 44.0 Gaussian Mixture ModelIn this report we implement the Gaussian Mixture Model for background subtraction. This background subtraction model is more robust than other models discussed in previous section. Most significantly it can handle multi-modal situations e.g. trees and sky which is more effectively filtered by the GMM model.Each pixel value is modelled by a categorization of Gaussian rather than a particular type of distribution. Based on the variance of each of the Gaussian of the mixture, we determine which Gaussians may correspond to background colors. Pixel values that do not fix the background distributions are considered foreground until at that place is a Gaussian that includes them with sufficient consistent evidence supporting it 5. This method is in truth adaptable with inflaming changes, repetitive motions and slow moving objects.This method contains 2 main significant parameters -Alpha, the learning constant and T, the proportion of the data that should be accounted for by the background The Gaussian mixture components for a pixel grant normalized weights calculated from the past observations. The likelihood that a pixel has a value of Xt is assumed to be 5The parameters of the mixture components are updates with the new frames. A thinkd pixel value is compared with all the components of the mixture assigned to that pixel to fall out out if there is a match. A match is said to recover when the retrieved pixel value is within 2.5 times standard deviation of a mixture component. The update procedure is different for the twinned component and other components. The mean values and the covariance matrices are updated for only the matching component. The value formulas for the matching com ponents areAnd the weights are updated with given formula 6In case there is not a match between the current pixel value and the mixture components related to that pixel .the component having the smallest likelihood with respect to the current pixel value is discarded. A new Gaussian component is created in place of the discarded one, having a mean value equal to the current pixel value, and a variance equal to a predetermined constant.To retrieve the foreground the following equation is used.Where, T is the threshold value for the sum of the weights.65.0 AlgorithmIn order to give a better discretion of the algorithm used for background subtraction the following steps were adopted to achieve the desired resultsFirstly, we compare each input pixels to the mean mu of the associated components. If the value of a pixel is close enough to a chosen components mean, wherefore that component is considered as the matched component. In order to be a matched component, the difference between the pixel and mean must be less than compared to the components standard deviation scaled by factor D in the algorithm.Secondly, update the Gaussian weight, mean and standard deviation (variance) to reflect the new obtained pixel value. In relation to non-matched components the weights w decreases whereas the mean and standard deviation chit the same. It is dependent upon the learning component p in relation to how debased they change.Thirdly, here we identify which components are parts of the background model. To do this a threshold value is applied to the component weights w.Fourthly, in the final step we determine the foreground pixels. Here the pixels that are identified as foreground dont match with any components determined to be the background.6.0 Experimental ResultsFor better results, median value filter was used where the main purpose was to filter out baffled pixels from the large connected pixels (e.g. people, vehicles) so that it was easy to distinguish between the background and foreground. With median filter the value of the output pixel is determined by the median of the neighboring pixels instead of the mean. Median has a much littler insensitivity compared with mean to extreme values.The function used was medfilt2(A,m n).Figure 3 GMM Background Subtraction with high light intensityIn the above image, the picture on the right shows the output of background subtraction using the GMM model. There is soothe false negatively charged foreground detection which produces the disturbances in the image. Although it still shows the objects moving hence, making it suitable for object tracking. The reason for these disturbances can be narrowed down to illumination changes.Figure 4 GMM Background Subtraction with Low light intensityFor the above, a short 30 second video was recorded where light intensity was moderate and the camera was still. It can be observed that the hand is visible and false negative detection is low which suggests that the mod el works robustly in low light intensities. From the above results, we can say that there are still improvements to be made in the algorithm mostly to work with illumination which increases the false negative background subtraction. 7.0 Further ImprovementsThe GMM model used in this paper could be further tweaked to provide better results in monetary value of zero disturbances/noises. The use of morphological filters could be implemented into the algorithm to reduce these unconnected pixels which were seen in the figures above and make it more robust in object tracking. The algorithm has the capabilities of further improvements to run large size videos and importantly the ability to process more frames per seconds using larger number of Gaussians in the mixture model. Apart from the above improvements, we further plan to research the model using PTZ (Pan Tilt Zoom) technique to study the processing rate of frame/second and observe the robustness of the algorithm in terms of disturb ances/noise. There are possibilities of false positive results where background could be recognised as foreground.8.0 ConclusionsIn this project we implemented a very strong and widely used background subtraction method according to the paper reconciling background mixture models for real-time tracking. This method is very adaptable for lighting changes and shadow removals. Also it finds the repetitive actions as well with the use of mixture of Gaussians. Alpha and T are the key parameters in this paper. The values of these parameters changes with the different cameras and different environment. So it is very important to get the best values for them to work with for different videos. Also we tried to improve the output type by using some filtration methods like median filter. The guassain model is not perfect with the result since there are some disturbances after the modelling also. But there are many good post processing techniques are introduced and we referenced two of them. A new region Gaussian background model for video surveillance by Xun Cai and Long Jiang, and ameliorate Post-Processing for GMM based Adaptive Background Modeling by Deniz Turdu,Hakan Erdogan.But the inaugural paper doesnt give any clear information about the techniques failed to explicate the equations as well. Second paper introduces a very good technique and we couldnt able to get the result properly. But this assignment was very intriguing for us and helped us to get a very good cognition about different background subtraction methods used in Video surveillance.9.0 References1 Robust techniques for background subtraction in urban traffic video Sen-Ching S. Cheung and Chandrika Kamath, 2006/072 Background Subtractions of Moving Objects https//computation.llnl.gov/casc/sapphire/background/background.html3 Page444, Computer vision A modern approach, David A. Forsyth4 Dr J. Zhang, CSC7006 Lecture 2 Slides, Queens University Belfast.5 Adaptive background mixture models for rea l-time tracking, Chris Stafer, W.E.L Grimson 6 Improved Post processing for GMM based adaptive background modelling by Deniz Turdu, Hakan Erdogan