Go to Polygence Scholars page
Aarush Sharma's cover illustration
Polygence Scholar2024
Aarush Sharma's profile

Aarush Sharma

Class of 2028Pleasanton, California



  • "How can supervised machine learning for image recognition be used in an attempt to strengthen perimeter security." with mentor Mitchell (Feb. 8, 2024)

Aarush's Symposium Presentation

Project Portfolio

How can supervised machine learning for image recognition be used in an attempt to strengthen perimeter security.

Started Sept. 19, 2023

Abstract or project description


With rising concerns about home break-ins in California, the necessity for robust perimeter security measures has become paramount. To address this issue, our project focuses on exploring various machine learning models and architectures to identify the most efficient model for object detection within the context of perimeter security. Through the collection of images from both front and back yard environments, we train neural networks to detect suspicious activities around the home's perimeter. The primary goal of this report is to emphasize the critical role of AI in bolstering home security measures. Through the implementation of facial recognition technology, the project aims to mitigate the possibility of crime by identifying unknown visitors at the door and promptly informing homeowners. Our methodology involves training models using Teachable Machine, a user-friendly platform for machine learning model development, and experimenting with Raspberry Pi for practical application. The discussions encompass the significance of facial recognition technology in enhancing security, ethical considerations surrounding its use, and suggestions for future improvements. While facial recognition technology offers promising benefits, it is crucial to address ethical, legal, and privacy concerns to ensure its responsible deployment. In conclusion, the integration of facial recognition technology into door cameras has the potential to significantly enhance home security, provided that it is implemented with careful consideration of its implications and continuous efforts to improve accuracy and fairness. Future endeavors should focus on expanding datasets and refining models to enhance accuracy and safety.