Head Count System Using Artificial Intelligence Method Information Technology Essay

Head Count System Using Artificial Intelligence Method Information Technology Essay Abstract This report is developing the project called Head Count System using Artificial Intelligence Method. This report compresses the introduction, literature review, methodology, project design, result and discussion, project management and conclusion of the project. For the starting of this report, there will shows more on the products (head count system) description, background, aims and objective. Next, this report will discuss more on the products component such as hardware and software. There will research or survey on the hardware and software those can be suit on designing the product. After the research or survey, this report will include details more into the specification of this project. Some discussion will also be done on the systems programming code, dimension of the hardware, and the linking tool and cable that are chosen to use. The key point of this project is on the artificial intelligence. Fuzzy logic, neural network, neurofuzzy, induction algorithms and genetic algorithms are the type of the artificial intelligence engine. This report will discuss on the structure, method and application behind the entire artificial intelligence engine in literature review. Furthermore, the detail discussion on neural network programming such as Hopfield network and support vector machine will carry out in the methodology. Some of the image processing programming will explained in that chapter too. After all the basic theory and methodology, the real work on designing this project will proceed in project design. There got three main programming those are image acquisition, network and counting system will elaborate in flow chart and figure. All the experiment result and discussion on this project will state in the following chapter too. Finally, the conclusion will conclude all the processes on research, design and building for this project in the end of the report. In the conclusion, the achievement, distribution and weakness of this project will stated out too. With the recommendation as the ending of this report, this project is hoping can be improved in future development. Introduction Head count system is using in everywhere and any time in the reality life. Counting the amount of head counts is a must since the world started no matter in which occasion such as lecture halls, parties, ballrooms and so on. Nowadays, there is still use a traditional ways to get the total amount of the head count. The ways are using such like count one by one manually by a person assigned; tick attendance using paper to record down and so on. Otherwise, some advance ways to get the head count such as using smart card to get the head count through a manually scanning device; manually check in through a control server and so on. Following the trend of the technology grown, a lot of automation products have been created or designed. These kinds of products are now available in the market to reduce the waste of time and energy on unnecessary purposes. If the head count system also can detect or check the amount of head count automatically, there will be increase the efficiency of work and prevent the resource is wasted. [1] Background of the project Head Count System using Artificial Intelligence Method is designed and developed head count system that is able to intelligently check the total number of head counts entering or exiting an enclosed location. This system can automatically determine the number of head count by using sensors and combination with artificial intelligence engine. So, this project will be built based on software and hardware too. Some more the interfacing device is needed for this project to link up the hardware direct to computer. This is because the artificial intelligence programming will write in software in computer and the input of sensor will direct get through the interfacing device. Methodology After the literature review on the project is done, the specification of the project is decided. This project will be use MATLAB as the programmer to write the code of Neural Network which is one of the AI engine, as the software part. Besides that, USB cable is chosen as the interfacing device to link the software and hardware. The hardware of this project is built by thermal imaging camera [13] Thermal Imaging Camera All materials, which are above 0 degrees Kelvin (-273 degrees C), emit infrared energy. The infrared energy emitted from the measured object is converted into an electrical signal by the imaging sensor (microbolometer) in the camera and displayed on a monitor as a color or monochrome thermal image. [23] Infrared is a form of electromagnetic radiation, radio waves, microwaves, ultraviolet, visible light, X-rays and gamma rays the same. All of these forms together constitute the electromagnetic spectrum, are similar, the speed of light they emit energy in the form of electromagnetic waves moving. Each other band in the spectrum of the main difference is their wavelength, and the related amount of wave energy to carry. For example, while gamma ray wavelengths than visible light millions of times smaller, the wavelength of radio waves over a billion times is visible. [23] With the technical explanations above, there show that thermal imaging camera is not affect by any light level and it is better than a normal camera by this point. Using thermal imaging camera in this project will reduce the faulty cause by light level to minimum. MATLAB MATLAB is a high-level language of technical computing. MATLAB can apply to wide application, including image processing, sound processing, financial analysis, control design, and data communications. In image processing, both 2-D and 3-D graphical image is able to function. Using MATLAB, technical computing problems can be solved faster than using traditional programming languages, such as C and C++. Besides that, inside MATLAB software contains a lot of useful toolbox. All the toolbox can be link up and function together. While programming in MATLAB, if any function command that not familiar can be get more information on that command by typing help followed with the command. With entire advantages shown above, MATLAB is chosen as the software platform for this project. [21] MATLAB is the platform using to test on all the sample method programming shown above. From the result on testing neural network sample programming, single-layer perceptron and multi-layer perceptron are good in use on data pattern classification. Hopfield network is suitable in image restore application. Then support vector machines and kernel function are used in pattern analysis especially in face detection system. For image processing sample programming, Kirsch edge detection and canny edge detection are good edge detecting solution. Shape recognition can be use Houdorff-based image comparison to implement. Gabor function can be applied with support vector machines for face detection. Lastly, one-dimensional Gaussion function is used in signal processing, but in image processing where two-dimensional Gaussian can be apply to it. After some preliminary work on so many programming, the Support Vector Machines has been chosen as the design programming on this project. The programming shown above is just the major of the programming linking with this project. The main reason chosen Support Vector Machines is this network combine both neural network and image processing method. Result and Discussion In this project result and discussion, the three main parts will be discussed in detail on the result. The three main parts are image acquisition able to get the input from sensor, train the network with support vector machines completely with no error and headcount system able count correctly. Figure 9: show the result on image acquisition For the image acquisition programming, it can test with PC camera. PC camera is set as the input in the programming. Right-hand side image on figure 9 at above shows the view point of the camera after the program is started. If there got any motion occurred in the view point of the camera, the image will be captured. Left-hand side image on figure 9 at above shows the result of captured image after motion occurred. The discussion above state all the basic function on this program and the result will show as figure in the real project. Figure 10: show the result on main program on creating network Figure 10 at above show the result after run the programming of support vector machines. The right-hand side of the figure 10 at above shows a prompt menu for the user chooses the step to perform on this program. If Generate Database is pressed, the program will load the image database to the workspace and there will shows Loading HeadCount and Loading non-HeadCount as shown at left-hand side, first two lines on figure 10. If the button Create SVM is pressed, the support vector machines network will be created. Some information of the network will be show out as shown at left-hand side, after first two lines on figure 10. To test the network is just press the Test on Photos button. After the button is pressed, another menu will prompt out is let the user to select the sample image for testing. After the sample is selected, the program will run automatically until the result come out. The program will circle all the headcount based on the network. If there got error, the program can l et the user to retrain it regarding to same sample. In the experiment on this project, one sample that contains some possibly of the combination and eight different orientation of headcount is created to use as the testing sample. Lastly, the menu will close if the Exit is pressed. Figure 11: show the result on GUI output for counting system For the counting system, GUI is created to show the result of the number of headcount as shown at right-hand side in figure 11 and the input image have been captured as shown at left-hand side in figure 11. The result is counted based on these theories: If the objects of first image touch the upper line then the objects of the following image touch the lower line, this mean the objects are moving in. If the objects of first image touch the lower line then the objects of the following image touch the upper line, this mean the objects are moving out. Conclusion This report shows the complete process on designing this project. After complete the literature review, this can help on choosing the hardware, software or interfacing device to use. Methodology of this project shows majority method or solution to do this project in both building hardware and the software programming. More testing and simulations are done to get better solution. The complete steps or procedures of this project are show in the project design with the flow chart. Practical testing is also done and the result has been show and discuss clearly. This project fully achieves the entire objective stated above. For the main objective on researching type of artificial intelligence engine has fully achieved. The definition, architecture, algorithm of all type of artificial intelligence such as fuzzy logic, neural network, neurofuzzy, induction algorithm and genetic algorithm have been explained and stated in the literature review. Summary in literature review support the point of achieve the main objective and prove the objective of determine the type of interfacing device to use for connecting hardware and software is achieved. By comparing the two types of the interfacing devices such as USB cable and RS232, USB cable has a better speed than RS232 cable. So USB cable is chosen. The investigating and testing the sample programming on both neural network and image processing in methodology is to achieve one of the objectives. That objective is investigated on the programming language for the artificial intelligence engine. There i s not only achieve the objective and some additional investigating on image processing is adding on. Design software and GUI which able to simulate the system model, and implement the prototype hardware to get the inputs are achieved objectives too. PC camera is used to get input with the worked programming and GUI indicate the number of headcount prove that these objective is achieved. The result is show and discussion clearly on above. The contributions and benefits of this project is save the life from emergency is happen, reduce the waste of time and energy on unnecessary purposes, increase the efficiency of work, and this project is ease to use head count system with fully automatically function. The weaknesses and disadvantages of this project is limit in light level on PC camera, time consuming and memory consuming on operate the system is high. Reference and Bibliography http://www.freepatentsonline.com/20090271240.pdf (Accessed: 30 October 2010) http://www.headcountsystems.com/counting_technology.html (Accessed: 9 October 2010) http://www.911headcount.com/solutions.html (Accessed: 9 October 2010) http://pragmatick.com/ (Accessed: 10 October 2010) http://www.sensourceinc.com/ (Accessed: 16 October 2010) http://www.countwise.com/products_overview.php (Accessed: 16 October 2010) http://www.infodev.ca/buildings.html (Accessed: 17 October 2010) http://www.trafsys.com/people-counters.aspx (Accessed: 17 October 2010) http://en.wikipedia.org (Accessed: 6 November 2010) http://www.wtec.org/loyola/kb/c1_s1.htm (Accessed: 6 November 2010) http://www.referenceforbusiness.com/management/A-Bud/Artificial-Intelligence.html (Accessed: 23 October 2010) http://www.zaptron.com/literature/neurofuzzy.htm (Accessed: 24 October 2010) http://www.mathworks.com/products/daq/supportedio14005.html (Accessed: 30 October 2010) http://www.automationdirect.com/static/specs/pe18mmfaled.pdf (Accessed: 31 October 2010) http://www.roboticsconnection.com/p-23-sharp-gp2d120-ir-sensor.aspx (Accessed: 31 October 2010) http://www.interfacebus.com/usb-cable-diagram-30.html (Accessed: 31 October 2010) http://faculty.olin.edu/bstorey/Notes/Card.pdf (Accessed: 30 October 2010) http://www.dataq.com/support/documentation/pdf/misc/MatlabAdaptor.pdf (Accessed: 30 October 2010) http://www.neuralwiki.org/index.php?title=Programming_the_USB-6008_with_Matlab (Accessed: 31 October 2010) http://en.wikipedia.org (Accessed: 31 October 2010) http://www.mathworks.com/products/matlab/description1.html (Accessed: 14 November 2010) http://www.developers.net/intelisdshowcase/view/47 (Accessed: 21 November 2010) http://www.inmes.hr/pdf/thermography.pdf (Accessed: 28 November 2010) http://www.phon.ucl.ac.uk/home/shl9/roel2a/smits2a.htm (Accessed: 5 December 2010) http://www.dtreg.com/mlfn.htm (Accessed: 12 December 2010) http://www.learnartificialneuralnetworks.com/hopfield.html (Accessed: 19 December 2010) http://www.dtreg.com/svm.htm (Accessed: 26 December 2010) http://crsouza.blogspot.com/2010/03/kernel-functions-for-machine-learning.html (Accessed: 26 December 2010) http://www.mathworks.com/products/image/ (Accessed: 2 January 2011) http://www.cvmt.dk/education/teaching/f09/VGIS8/AIP/canny_09gr820.pdf (Accessed: 2 January 2011) http://mathworld.wolfram.com/GaussianFunction.html (Accessed: 9 January 2011) http://www.pages.drexel.edu/~weg22/can_tut.html (Accessed: 16 January 2011) https://ece.uwaterloo.ca/~ece327/old/2007t1/proj/doc/proj.pdf (Accessed: 23 January 2011) http://www.cs.cornell.edu/~dph/hausdorff/hausdorff1.html (Accessed: 30 January 2011) http://www.cs.cornell.edu/vision/hausdorff/hausmatch.html (Accessed: 13 February 2011) http://www.cnbc.cmu.edu/~tai/papers/pami.pdf (Accessed: 20 February 2011)