INTRUDER ALERT SYSTEM

RAM REDDY RAMYA REDDY KAVYESH PASHAM

HISTORY: GUN VIOLENCE IN US SCHOOLS

Gun violence manifests in a myriad of ways in American schools, and school shootings have created new anxieties for the younger generation of students. According to an Everytown analysis, there have been at least 405 incidents of gunfire on school grounds from 2013 to 2018. Of these, 260 occurred on the grounds of an elementary, middle, or high school, resulting in 109 deaths and 219 injuries. While mass shootings like the incident at Sandy Hook and, more recently, Parkland and Santa Fe are not commonplace, schools are more likely to experience homicides and assaults, unintentional shootings resulting in injury or death, and suicide and self-harm injuries. All incidents of gun violence in schools, regardless of their intent or victim count, compromise the safety of students and staff.

PURPOSE • There have been a lot of gun-violence related activities that are caused in schools by ex-students/unknown intruders that enter the school campus. To address these ever present issues in society, This project helps schools in identifying intruders when they attempt to enter schools, by using facial recognition software that gives an alert to the school staff. Since schools take pictures of all their students when making student IDs, it is assumed these pictures are stored in a database, and can thus be used to analyze the faces of the persons who go into a school. • Engineering Goal for this project is identifying the intruder by scanning the facial image of each student and matches with the back end database. If any student facial image does not match with the existing images of the student database gives alert to the office staff. • Resources: OpenCV, Python software and pycharm, webcam, laptop

WHAT IS FACIAL RECOGNITION?

Facial recognition systems are computer based security systems that are able to automatically detect and identify human faces

based on the facial characteristics and features derived from a camera or digital

photo.

The development stage for facial recognition began in the late 1980s and were commercially available by the 1990s. While many people first heard about facial recognition after September 11th, 2001.

SYSTEM OVERVIEW

The implementation of face recognition technology includes the following stages.

Image acquisition

Image processing Distinctive character

location

Template creation Template matching

TRAINING DATA STRUCTURE

The students photos are stored in the training-data folder. Each student photos are kept in s1, s2, .....sn subfolders as shown in the hierarchy. Multiple photos of different poses of each student are kept for training data for facial matching.

training-data | | -- s1 | |-- 1.jpg | |-- .... | |-- 30. jpg | -- s2 | |-- 1. jpg | |-- ..... | |-- 30.jpg ..... | -- s40 | |-- 1.jpg | |-- .... | |-- 30.jpg

IMPLEMENTATION

The implementation of face recognition technology includes the following fourstages. • Capture: An image sample is captured by the system • Extraction: Face is detected and a template is created from the extracted data • Comparison: The template is then compared with a new sample. • Match/Non-Match: The system decides if the features extracted from the new samples are a match or a non-match

ANALYSIS OF DATA - 1

0 20 40 60 80 100 120

2 4 6 8 10 12

Accuracy (percent)

Distance (ft)

The Effect of Distance(ft) on Facial Detection Accuracy(%) of Various Cameras

1080P 720P Distance( ft )

Percent (1080P)

percent (720P) 2 100 100 4 98 96 6 93 89 8 87 74 10 76 62 12 62 49

ANALYSIS OF DATA – 2

0 20 40 60 80 100 120

10 20 30 40 50

Percent

No. Training Data Images

The Effect of No. Images on Facial Detection Accuracy(%) of Various Cameras 1080P 720P

No. Images

Percent (1080P)

percent (720P) 10 71 65

20 82 79

30 92 90

40 97 94

50 100 100

STATISTICAL ANALYSIS OF CAMERA QUALITY

Distance on Accuracy

• H0 : μDifference = 0 HA : μDifference > 0

•t = 2.959 • P=0.016, α = 0.05, p < α • Significant difference between camera quality

No. Images on Accuracy

• H0 : μDifference = 0 HA : μDifference > 0

•t = 2.89 • P=0.022 , α = 0.05, p < α • Significant difference between camera quality

ADVANTAGES AND LIMITATIONS

• Non-intrusive: can even be used without subjects knowledge. Other biometrics require subject co-operation and awareness Examples: • Iris recognition – looking into the eyes • Fingerprint – placing hand on fingerprint reader

• Human readable media and can be verified by a human • No association with crime, as with fingerprints • Data required is easily obtained and readily available

• Many different views per person are needed in the database • No lighting variations or facial expressions are allowed • High computational cost due to iterative searching

CONSIDERATIONS FOR A FACIAL RECGNITION SYSTEM

SIZE AND QUALITY OF DATABASE

LIGHTING CONDITIONS

QUALITY OF THE CAMERA

USER BEHAVIOR

HOW LONG SINCE LAST IMAGE ENROLLED

REQUIRED THROUGHPUT RATE

MINIMUM ACCURACY REQUIREMENTS

MODE OF OPERATION

APPLICATIONS FOR FACIAL RECOGNITION TECHNOLOGY

• Government Use • Law Enforcement • Counter Terrorism •Immigration • Voter Verification •Issuance of Drivers License • School Security

• Commercial Use • Banking - ATM • Residential Security • Day Care • Gaming Industry

OPEN RESEARCH PROBLEMS

NO GENERAL SOLUTION FOR VARIATIONS IN FACE IMAGES LIKE ILLUMINATION AND POSE PROBLEMS

PROBLEM OF AGEING

CONCLUSION

Our major finding for this experiment was the effect of camera quality, distance, and number of images in the database on the accuracy of facial detection. Camera quality and number of images in the database positively affected facial detection accuracy, while distance has a negative impact. To further improve this experiment, it would be optimal to test accuracy in various lighting conditions as well as more high quality cameras.

REFERENCES

• XPERT OPINION: THE EFFECT OF ARTIFICIAL INTELLIGENCE ON FACE RECOGNITION. (n.d.). Retrieved

January 4, 2020, from Everteam website: https://www.everteam.com/en/artificial-intelligence-face- recognition/

• Ivancic, K. (2019, February 27). Traditional Face Detection With Python. Retrieved January 4, 2020, from Real Python website: https://realpython.com/traditional-face-detection-python/ • Tiwari, S. (n.d.). Face Recognition with Python, in Under 25 Lines of Code. Retrieved January 4, 2020, from Real Python website: https://realpython.com/face-recognition-with-python/ • Dwivedi, D. (2018, April 28). Face Recognition for Beginners. Retrieved January 4, 2020, from Medium website: http://Face Recognition for Beginners • Balamurugan, S. (n.d.). Face Recognition Using LBPH Algorithm. AI Time Journal. Retrieved from https://www.aitimejournal.com/@shanmugapriya.balamurugan/face-recognition-using-lbph-algorithm

Built With

Share this project:

Updates