Face Anti Spoofing Through Raspberry Pi Using Machine Learning

Project Title

Face Anti Spoofing Through Raspberry Pi Using Machine Learning

Project Area of Specialization Information & Communication TechnologyProject Summary

Due to advances in technology, the face recognition system has become an indispensable component in numerous real-world applications requiring secure verification or recognition, such as military, banking, surveillance, smart houses, and device unlocking, because of its high level of security and reliability. Since there is no contact required for facial recognition like there is with fingerprinting or other security measures, facial recognition offers a quick, automatic, and seamless verification experience. There is nothing such as a key or I.D that can be lost or stolen.

Unfortunately, face biometric system is vulnerable to spoofing attacks where spoofing can be performed using a substitute for another’s person’s face – usually their photo, facial cosmetic makeup, video recording, 3D mask. etc. If the spoofing attack succeeds, the fraudster acquires privileges or access rights of another person. To prevent face recognition systems from the criminals, face anti-spoofing technology is introduced that distinguishes between real individuals and face spoofs.

Project Objectives

My undergraduate project has been mainly performed within the context of the project with a focus on improving the security of face biometric systems. More specifically, I have completed discussion with my current mentor and done literature review relevant to face anti-spoofing techniques with the help of my supervisor to achieve the targets on time. Hence, our main objectives are listed below:

1. Implement a deep learning method for traditional spoofing attacks (prints or videos) to detect a variety of 2D attacks.
2. Analyze the performance with traditional machine learning algorithms such as SURF, KAZE, etc. and currently published state-of-the-art methods.
3. And also use some alogorithms of CNN, Residual Network, Liveness Factor.
4.  Visual representation with raspberrypi.

Project Implementation Method

Benefits of the Project

According to the research, some countries showed strong improvement in face anti spoofing in the past few years. But some of these researchers of the countries in 2018 to 2020, are work on only 3D photos/video detection and in 3D using PAD,CNN features and RGB color space feature.All these researcher didn’t work using some other features of machine learning. So in this Project we will use Some deep learning and add some more features including algorithem like SURF, KAZE , figen Faces etc. we aslo use Raspberry Pi with  machine learning and connect the web camera to show the effective result.

Technical Details of Final Deliverable

The final deliverable sysyem is capable of detecting the face without being spoof by criminals. We are doing some machine learning on rasberry pi including surp and kaze algorithums which differniate this model from local face recognition system. So we are commercially deploying our project within our university. We are presented this project which is more secure than local face recognition system.

Final Deliverable of the Project HW/SW integrated systemCore Industry SecurityOther Industries IT Core Technology OthersOther Technologies 3D/4D Printing, Clean TechSustainable Development Goals Quality EducationRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	44000
rasberry pi	Equipment	2	9000	18000
web camera	Equipment	2	4000	8000
miscellaneous items	Miscellaneous	2	4000	8000
GPU based computer	Equipment	2	5000	10000