ABSTRACT:
The proposed system is a low-cost AI-based vision application developed to perform both real-time face recognition and object detection using computer vision and deep learning techniques on a Raspberry Pi platform. The system is implemented in Python and integrates multiple libraries such as OpenCV, NumPy, Tkinter, Picamera2, and Ultralytics YOLOv8 to achieve efficient image processing and user interaction in an embedded environment.
The first module of the system focuses on face dataset creation and recognition. Using the Raspberry Pi Camera Module (Pi Camera), the system captures facial images through live video streaming. Haar Cascade classifiers are employed to detect human faces from the video frames. The detected face regions are then cropped, converted to grayscale, resized, and stored in a structured dataset directory. This dataset is used to train a face recognition model based on the Local Binary Pattern Histogram (LBPH) algorithm, which is well-suited for low-power devices like Raspberry Pi. During real-time recognition, the system compares detected faces with the trained model and identifies known individuals. If a face does not meet the confidence threshold, it is labeled as “Unknown”, and the corresponding face region is automatically blurred to ensure privacy and security.
The second module integrates YOLOv8 (You Only Look Once version 8), a state-of-the-art deep learning model for real-time object detection. The model processes video frames captured from the Pi Camera and detects multiple real-world objects simultaneously, drawing bounding boxes and class labels with high accuracy and speed. This enables the system to recognize common objects, people, and environmental elements effectively.
A Graphical User Interface (GUI) is developed using Tkinter to provide a simple and user-friendly control panel. The interface includes options to capture face datasets, train the recognition model, perform face recognition, run object detection, and safely exit the application. The system also incorporates efficient camera management, error handling, and structured data storage to ensure reliable performance on Raspberry Pi hardware.
Overall, the developed application functions as a compact and intelligent AI vision system that combines biometric authentication and object detection into a single platform. Due to its low cost, portability, and real-time processing capabilities, the system is suitable for applications such as smart surveillance, access control, attendance monitoring, and embedded AI-based security systems.


INTRODUCTION:
In recent years, the rapid advancement of Artificial Intelligence (AI) and Computer Vision has enabled machines to interpret and analyze visual data with high accuracy. These technologies are widely used in applications such as surveillance systems, biometric authentication, smart security, and real-time monitoring. Among these, face recognition and object detection are two important areas that play a crucial role in enhancing automation and security in modern systems.
Face recognition is a biometric technology that identifies or verifies a person based on facial features. It is commonly used in access control systems, attendance monitoring, and identity verification. On the other hand, object detection is a deep learning-based technique that allows systems to identify and locate multiple objects in real time, making it useful for applications such as smart surveillance, autonomous systems, and environment monitoring.
This project aims to develop a low-cost and efficient AI-based vision system that integrates both face recognition and object detection into a single platform. The system is implemented using Python and deployed on a Raspberry Pi, making it portable and suitable for embedded applications. A Pi Camera module is used to capture real-time video input, which is processed using computer vision algorithms.
For face recognition, the system uses Haar Cascade classifiers for face detection and the Local Binary Pattern Histogram (LBPH) algorithm for training and recognizing faces. The system is capable of identifying known individuals and labeling unknown persons, with an added feature of blurring unknown faces to enhance privacy and security. For object detection, the system utilizes the YOLOv8 deep learning model, which provides fast and accurate detection of multiple real-world objects in real time.
To improve usability, a Graphical User Interface (GUI) is developed using Tkinter, allowing users to easily interact with the system through options such as dataset capture, model training, face recognition, and object detection. The integration of these technologies results in a compact and intelligent vision system that can be used in various real-world applications.
Overall, this project demonstrates how AI and embedded systems can be combined to create a smart, efficient, and cost-effective vision-based solution for modern security and monitoring needs.

OBJECTIVES:
The main objectives of the proposed Real-Time Object Detection and Masking Framework are:
1. To develop a real-time vision-based detection system
o Detect human faces and multiple objects using live camera input.
2. To implement accurate face detection and recognition
o Use computer vision techniques to identify known individuals from a trained dataset.
3. To build a reliable face dataset collection module
o Capture and store user images in a structured format for model training.
4. To apply an efficient recognition algorithm
o Utilize the LBPH (Local Binary Pattern Histogram) algorithm for fast and lightweight face recognition.
5. To identify and handle unknown individuals
o Detect unrecognized faces and label them as “Unknown” in real time.
6. To implement automatic face masking (blurring)
o Apply blur to unknown faces to ensure privacy protection and security.
7. To integrate real-time object detection
o Use YOLOv8 to detect and classify various real-world objects with bounding boxes.


8. To ensure smooth performance on embedded systems
o Optimize the system to run efficiently on Raspberry Pi with Pi Camera.
9. To design an interactive graphical user interface
o Provide easy-to-use controls for dataset capture, model training, recognition, and detection.
10. To develop a compact and low-cost AI solution
o Create a portable system suitable for smart surveillance, access control, and monitoring applications.