Computer Vision Projects — OpenCV & Python

A collection of computer vision projects built with Python and OpenCV, covering real-time face detection, face recognition, motion-based security systems, and an AI voice assistant. All projects run locally using a webcam or image input.

---

Projects

1. Real-Time Face Detection (alert.py, Face Recognition.py, import cv2.py)

Real-time face detection from a live webcam feed using OpenCV's Haar cascade classifier.

• Captures live video stream
• Detects faces frame-by-frame using haarcascade_frontalface_default.xml
• Draws bounding rectangles around detected faces
• Prints detection alerts to console

---

2. Face Recognition System (face_train.py, face_detect.py)

A trained face recognition model that identifies individuals from images or live input.

Training (face_train.py):

• Loads images from the images/ folder (organised by person)
• Extracts facial features using OpenCV
• Trains a recognition model and saves it to face_trained.yml

Recognition (face_detect.py):

• Loads the trained model
• Detects and identifies a face in a test image
• Returns the predicted label and confidence score

Subjects trained on: Messi, Elon, Bezos, Ryan Reynolds, Canu

---

3. Motion Detection Security Camera (newsecurity.py, security.py)

A motion-sensitive security system that triggers an alert when movement is detected.

• Compares consecutive webcam frames to detect movement
• Triggers a siren audio file (siren.wav) or system beep on motion detection
• Runs continuously until manually stopped

---

4. AI Voice Assistant (New Ai.py)

A voice-activated assistant using Vosk for offline speech recognition.

• Listens for voice commands via microphone
• Handles commands such as playing YouTube videos and fetching Wikipedia summaries
• Uses pyttsx3 for text-to-speech responses

---

5. Intent Classifier (model_train.py, intent.json)

Prepares training data for an NLP intent classification model.

• Loads intent patterns from intent.json
• Processes and labels patterns for machine learning training

---

Tech Stack

Area	Libraries
Computer Vision	OpenCV (cv2)
Face Recognition	OpenCV LBPH Recognizer
Speech Recognition	Vosk
Text-to-Speech	pyttsx3
Audio	simpleaudio, winsound
Web Automation	pywhatkit
Information Retrieval	wikipedia
Data Processing	NumPy, scikit-learn

---

Getting Started

Install dependencies

pip install opencv-python numpy scikit-learn vosk pyttsx3 simpleaudio pywhatkit wikipedia

Train the face recognition model

python face_train.py

Run face detection

python face_detect.py

Run the security camera

python newsecurity.py

Run the voice assistant

python "New Ai.py"

---

Project Structure

computer_vision/
├── images/                             # Training images organised by person
│   ├── Messi/
│   ├── Elon/
│   ├── Bezoz/
│   ├── Ryan/
│   └── Canu/
├── Face Recognition.py                 # Real-time face detection (webcam)
├── New Ai.py                           # Voice assistant
├── alert.py                            # Face detection with console alerts
├── face_detect.py                      # Face recognition inference
├── face_train.py                       # Face recognition model training
├── haarcascade_frontalface_default.xml # Haar cascade classifier
├── import cv2.py                       # Basic face detection script
├── intent.json                         # Intent patterns for NLP training
├── model_train.py                      # Intent classifier data preparation
├── newsecurity.py                      # Motion detection with audio alert
└── security.py                         # Motion detection with system beep

---

Note on YOLOv8 Object Detection

A separate real-time object detection project using YOLOv8 and the COCO dataset was developed in Google Colab. It covers custom model training, multi-object recognition across diverse categories, and live webcam inference. The Colab notebook will be linked here once published.

---

Author

Olaniyan Taslim — github.com/Asiwaju24