Face Recognition as Security for Door Accessing using Raspberry Pi

Face Recognition as Security for Door Accessing using Raspberry Pi

P Roja Kamala, Dr. D. Ravi Krishna Reddy Gudlavalleru Engineering College, Gudlavalleru, Krishna (Dist.), Andhra Pradesh.

Abstract: The embedded system performs some of the image processing tasks and sends the processed data to a PC. Using raspberry pi the PC tracks persons and recognizes two-person interactions by using camera. Based on our previous research, we explored the optimum division of tasks between the embedded system and the PC. Image Processing implemented in hardware have emerged as the most viable solution for improving the performance of image processing systems using cameras. Tracking, recognizing and detecting objects using image or video sequence are topics of significant interest in computer vision. To overcome this problem, we propose an embedded system for each camera that performs the fundamental image processing tasks (that is, the human segmentation processing) and sends the output data to a PC through GSM. The system consists of a single video camera, a video- capturing device and a PC.

Keywords : Raspberry pi, camera, personal computer, GSM..

INTRODUCTION

The technology wants to develop an amazing attentive technical innovation nowadays. Deep Learning is one in every of the fascinating area that permits machine to self-coach by providing input of datasets and gives a proper data during tests by applying many learning calculations. Recently Attendance is taken into account a crucial factor for both thecode likewise becausethe teacherof an academic organization. Maintaining attendance in educational institutions is incredibly critical. Each institution has its way of taking student attendance. In real time , the attendance management of the students can be managed in two different forms namely, MAS and AAS.MAS Management system could be a procedure where a tutor troubled with the actual fact must call the scholar and mark the attendance manually.

So, the matter emerges once someone give some thought to the normal routine of marking attendance within the classroom.

To remove these problems we keep company with Automated Attendance System(AAS).It happens by identification of faces using the literal unique characteristics of their faces. Facial recognition has the most unique characteristics that all other biometrics don't have.

Facial images may be taken from a distance and for authentication, no special action is required.

• Observations

  The face recognition technology is widely applied, not only to security applications but also to image indexing, image recovery, and natural user interfaces. Faces are extremely complex and dynamic objects that are used as biometric proof for biometric identification. the performance of face identification methods has been highlighted and CC is one of the foremost promising technology paradigms that may be wont to achieve it. Face recognition software was also utilized in a classroom to test and recognize students. The attendance of the understudies might be fixed on by catching their appearances in a top-notch video spilling administration, so it turns out to be profoundly reliable to the machine to know the presence of the researchers inside the study hall. while the brilliance-based methodology likewise named on account of the worldwide face acknowledgment framework, used in perceiving all the pieces of the picture.

• Literature Survey

  Using the AI approach a frame work is designed and executed to take the feedback and the attendance in the paper [1] are the two enhancements done for students. This technique distinguishes the codedexecution naturally and keeps up records like participation and criticism on subjects like science, English, and so on, during this way the feedback and the attendance is taken by perceiving the face of the that particular student, if the face is already registered then the student can access the portal. Also, the students can access other records. In the traditional system username and passwords will be given, which can be easily shared to the others, which can be big disadvantage.

  The category which the program will check the participation by remembering him/her. This Algorithm has been used for face recognition which recognizes external body part utilizing course classifier and PCA calculation for include determination and SVM for characterization. The gadget spares time when contrasted with customary participation markers and furthermore helps track the understudies.[2]

  In this paper Right now, the understudy is mentioned to remain before the camera to distinguish and perceive the iris, for the framework to check participation for the understudy. A few calculations like Conversion, Six Segment Rectangular Filter, Skin Pixel Detection is being used to recognize the iris. It helps in forestalling the intermediary problems and it keeps up the participation of the understudy in a compelling way, yet in one of the tedious procedures for an understudy or a staff to hold up until past individuals are done.[3]

  This paper recommends that the framework takes the participation naturally acknowledgment acquired by consistent perception. Constant perception helps in evaluating and updating the exhibition of participation. This project is focused round the methodology to encourage the changed a great deal of each connected with seat as demonstrated by its region. The viability of the image is likewise being examined to empower the quicker acknowledgment of the image.[4]

  AUTOMATED ATTENDANCE SYSTEM USING MACHINE LEARNING APPROACH:

  In this paper the authors proposed a technique that is customary for grading the participation by the instructor that requires the more time and extensive pressure that includes the intermediary and blunders participation. The understudys quantities are expanding step by step for all the schools, colleges, offices and other institutions across to keep the records. In the present scenario all the frame works are computerized by utilizing the iris (EYE SCANNERS), bio metric (FINGER PRINTS) are used in the huge place and for different purposes. These systems give the acknowledgements to the user. If any unauthorized person was detected then it give an message on the device screen that not detected but it does not send the message to the administrator. To beat these issues, biometric highlight like facial acknowledgment can be utilized which includes the stages, for example, picture procurement, face identification, highlight extraction, face grouping, face acknowledgment and inevitably denoting the participation. The calculations like ViolaJones and HOG includes alongside SVM classifier are utilized to obtain the ideal outcomes. Different ongoing situations should be viewed as, for example, scaling, light, impediments, and posture. The issue of excess in manual records and keeping participation is comprehended by this framework. Quantitative investigation is done based on PSNR values [5]

  PANORAMIC FACE RECOGNITION:

  Regularly, displaying 3D face is a powerful path for present invariant acknowledgment, yet its costly calculation essentially debilitates expected applications. In this paper, a basic and completely programmed all-encompassingpicture-based poseinvariant face acknowledgment strategy is proposed to give amazing precision low unpredictability. In this paper, a face shape model with neighbourhood transforming treatment is first built and considered as the arrangement standard to manage the entirety of the conceivable geometric bending issues. During the acknowledgment stage, a proposed efficiently structure calculation with transforming and the choice capacity are both used to fundamentally facilitate the negative impacts of different stances inside ±45° in yaw and ±22.5° in pitch. As showed in exploratory outcomes, a comparative precision as that of the 3D beginning of human expressions is accomplished with considerably less computational complexity [6]

  MOTION BASEDATTENDANCE SYSTEM IN REAL TIME ENVIRONMENT FOR MULTIMEDIA APPLICATION :

  Present day homerooms are outfitted with different savvy gadgets that have supporting programming to improve the study halls. In any case, it is frequently observed that the valuable time of homeroom is squandered on gauging participation, or the class may confront different interference because generally sections of understudies, and this participation method is done more than once. Likewise, execution is influenced by the nearness of the understudy in the organization. Ordinary strategies for participation taking are by calling understudy name or by sensor based card (RFID sensor) or biometric unique mark based participation framework, yet these technique are wasteful as should be obvious whether the understudy has gone to full class or not. In this way there is need of participation the executives framework which is thoroughly movement based that will help the resources and foundation for denoting the nearness of the understudy participation. Movement based participation framework that help recording and the executives of participation that will be utilized for different organizations and the colleges.[7]

  AN IMPLEMENTATION OF FEATURE EXTRACTION OVER MEDICAL IMAGES ON OPEN CV ENVIRONMENT:

  Highlight extraction is a procedure of recognizing and extricating highlights out of pictures and putting away it in include vectors. It is a significant period of substance-based picture recovery. Separated highlights are similarly noteworthy regarding their use since they can be additionally taken as contributions to next periods of CBIR. There are a few procedures utilized for separating highlights out of a picture. In this paper we show highlight extraction taking a shot at clinical pictures utilizing a SURF procedure under the OPEN CV stage. We have utilized the OPEN CV stage since it gives a C interface to actualize different picture preparing calculations. The work combines the element extraction method with this most appropriate stage accessible for picture calculations. We have additionally figured the presentation of the method utilized as far as different boundaries like execution time, turn, identify capacity, exactness, and so on. Our strategy functions admirably regarding its exhibition as appeared in our outcomes thus it very well may be one of the proper strategies to be picked for clinical pictures being the complex images [8]

• Modules used

1. Open CV

   In the year 1999 intel as started an Open CV that will be notable by the personal competer by using the vision library. The cross-arrange library sets its consideration on progressing picture getting ready and fuses sans patent executions of the latest PC vision computations.

   It reveals to you the most ideal approach to perform face affirmation with FaceRecognizer in OpenCV (with full source code postings) and gives you an introduction into the computations behind. I'll similarly advise the most ideal approach to cause the recognitions you to can find in various preparations considering the way that numerous people mentioned it.

   The presently available methods :

   Eigenfaces (see createEigenFaceRecognizer) Fisherfaces (see createFisherFaceRecognizer)

   Local Binary Patterns Histograms (see createLBPHFaceRecognizer)

2. FACE -DETECTION:

   Face detection is the primary stage in the ID procedure. Face location is known to be a class of two (face versus nonface). Identifying faces from a picture, video following, might be exposed to different restrictions, for example, enlightenment issues, present varieties, impediments because of adornments, and so on. Different methodologies like all- encompassing/layout, geometric formed were utilized to address the constraints that were expressed. The strategies utilized had the option to determine just a couple of restrictions.

   Utilizing a blend of systems promising outcomes were seen. The face's geometric highlights are gotten from the ears, mouth structure, face outskirt, and so on and are masterminded as a displaying and acknowledgment map The different face recognition systems are Principal Component Analysis (PCA)

3. FACE MODELING

   Face demonstrating produces an individual's prepared face model and stores it for approval in the database. Demonstrating faces is a vital piece of facial acknowledgment. To offer a realistic face template, the facial form is combined with distinctive facial features. Face demonstrating clarify the definition of face caluculation complexity. The face demonstrating process also has elastic bunch graph, active appearance model , etc.

4. FEATURE EXTRACTION

   Extracting the face expression is a necessary procedure associated with face location, face displaying, face Recognition, movement, assurance of face appearance, model-based coding of pictures, and so forth. Extraction of the facial quality is very sensitive to noise, shifts in lighting and pose. Geometric based, shading division based, appearance-based, format-based strategies are wont to expand the exhibition of the face acknowledgment framework. inside the programmed face acknowledgment framework, the face picture is distinguished, highlights are separated, and recognized. The foremost well- known methodologies utilized are graphical based strategies, format-based systems, appearance-based procedures, and so on. For feature extraction, face geometric configuration is considered in geometric techniques. Face expression like ears, nose,

   mouth, eyebrows, face shape, and their spatial relationship are wont to derive characteristics. Edge recognition techniques, slope examination strategy, hear like element hinder in Ada help strategy, and so on are utilized. Layout based strategies remove face demeanor like ears, nose, mouth, and so forth, supported the structure of the template and therefore the correct energy balance. for instance, the format-based component deriving strategy doesn't speak to a worldwide face structure while the appearance-based element extraction technique represents a worldwide face structure with a high computational expense.

   Local Binary Patterns Histograms (LBPH)

   Local Binary Pattern (LBP is a basic yet productive surface administrator which marks the pixels of a picture by thresholding the area of every pixel and thinks about the outcome as a parallel number

   the LBPH has 4 parameters: Radius ,Neighbors ,Grid X ,Grid Y

   • Implementation of the Proposed model

Figure : Block diagram of Proposed system

Figure : Hardware implementation of the proposed system.

In the above diagrams we observe that the Raspberry Pi Board is connected with a camera is place for the face detection of the user. A buzzer is placed if any unknown person tries to open the door, immediately the authorised person will get an warning message.

Flowchart 1 :Flowchart for dataset process

The task of the recommended structure is to get the quintessence of each understudy and to fix it inside the database for their support. The substance of the understudy ought to be trapped in such some manner, that each one the component of the understudies' face ought to be recognized, even the seating and thusly the position of the understudy ought to be seen.

There's no prerequisite for the teacher to truly measure support inside the class considering the established truth that the system information a video and through further taking care of advances the face is being seen and along these lines the interest information base is revived.

Flowchart 2 : Procss of Recognition

Flowchart 3:Face recognition

EXPERIMENTAL RESULTS:

Figure 1 : Recognition of authorized person

Initially all the images are registered in the data base now when the user come in front of the camera. It takes the picture of the person and matches with the images in the data base. If the picture is not matched with the available image in the data base then it displays the name of the person , if it not available in the data base then it shows unknown, as shown in the above figure above .

In the next step the proposed system also sends the messages to the registered end user where the persons/ users are the authorized or not. The messages/notifications will be received in the below given format.

Figure 2: Unauthorized person detection

Figure: message send to the authorized end user

CONCLUSION

We structured System which diminishes human endeavors and provides security. The proposed framework is modest, dependable and segments are effectively accessible. it's likewise versatile and effectively upgradable. The framework gives Security locks to entryway, solace, security and vitality effectiveness for the client. Raspberry Pi works and controls movement

finders and cameras for catching the image of the visitor. the longer term extent of this task is to feature a further module to the progressing venture which is that the GSM module, with the help of this GSM module we are able to send an alarm message as far as a secret phrase.

REFERENCES

1. N.H. Ismail, Zarina Tukiran,N.N. Shamsuddin, Faculty of Electrical and Electronic Engineering University Tun Hussein Onn Malaysia Johor, Malaysia

2. Raj Kamal, "Embedded Systems – Introductions", Publications:McGraw-HillEducation.Wi-Fitechnology, Technology Tracking Department ,2003

3. A. Kassem, S. E. Murr, G. Jamous, E. Saad, and M. Geagea, A smart lock system using Wi-Fi security, 2016, pp. 222225.

4. Jain, Ankit, Anita Shukla, and RituRajan, Password Protected Home Automation System with Automatic Door Lock, 2016.

5. Rathod, H., Ware, Y., Sane, S., Raulo, S., Pakhare, V., & Rizvi, I. A. (2017, January). Automated attendance system using machine learning approach. In 2017 International Conference on Nascent Technologies in Engineering (ICNTE) (pp. 1-5). IEEE.

6. Liu, Y. F., Guo, J. M., Liu, P. H., Lee, J. D., & Yao, C. C. (2017). Panoramic face recognition. IEEE Transactions on Circuits and Systems for Video Technology, 28(8), 1864-1874.

7. Yadav, M., Aggarwal, A., & Rakesh, N. (2018, January). Motion Based Attendence System in Real-Time Environment for Multimedia Application. In 2018 8th International Conference on Cloud Computing, Data Science & Engineering (Confluence) (pp. 332-336). IEEE.

8. Kaur, B., & Jindal, S. (2014, September). An implementation of Feature Extraction over medical images on OPEN CV Environment. In 2014 International Conference on Devices, Circuits and Communications (ICDCCom) (pp. 1-6). IEEE.

9. C.-H. Hung, Y.-W. Bai, and J.-H. Ren, Design and implementation of a single button operation for a door lock control system based on a near field communication of a smartphone, 2015.

10. J. Potts and S. Sukittanon, Exploiting Bluetooth on Android mobile devices for home security application, 2012.

11. Rathod, H., Ware, Y., Sane, S., Raulo, S., Pakhare, V., & Rizvi, I. A. (2017, January). Automated attendance system using machine learning approach. In 2017 International Conference on Nascent Technologies in Engineering (ICNTE) (pp. 1-5). IEEE.