Niranjan1005/FaceRecognition

---

Face Recognition System Documentation

Introduction

This document provides an overview and usage instructions for a face recognition system implemented in Python. The system consists of three main components:

1. Face Dataset Creation: Captures facial images and organizes them into a dataset.
2. Recognizer Training: Trains a face recognizer using the created dataset.
3. Real-time Face Recognition: Utilizes the trained recognizer to perform face recognition in real-time.

Components

1. Face Dataset Creation

Class: CreateDataset

• Attributes:

  • face_cascade_path: Path to the Haar Cascade XML file for face detection.
  • face_id: Initial face ID for labeling captured images.
  • face_images_per_id: Number of face images to capture per ID.
  • face_counter: Initial face counter.
  • dataset_folder: Path to the main dataset folder.

• Methods:

  • detect_faces(img, gray): Detects faces in a given image.
  • draw_rectangles_and_save_images(img, gray, faces, subfolder): Draws rectangles around detected faces, saves images, and displays them.
  • start_capturing(): Initiates the face capturing process.

Usage:

face_dataset_creator = CreateDataset(
    face_id=args.face_id,
    face_images_per_id=args.images_per_id,
    face_counter=args.face_counter,
    dataset_folder=args.dataset_path,
)

face_dataset_creator.start_capturing()

2. Recognizer Training

Class: FaceTrainer

• Attributes:

  • data_path: Path to the dataset folder.
  • trainer_folder: Path to the folder where the trained recognizer will be saved.
  • cascade_path: Path to the Haar Cascade XML file for face detection.
  • recognizer: LBPH Face Recognizer.
  • detector: Cascade Classifier for face detection.

• Methods:

  • _get_image_id(image_path): Extracts the face ID from an image path.
  • _get_images_and_labels(): Processes images from the dataset and extracts face samples and corresponding IDs.
  • train_recognizer(): Trains the LBPH Face Recognizer.

Usage:

face_trainer = FaceTrainer(data_path=args.dataset_path)
if face_trainer.train_recognizer():
    print("Recognizer trained successfully!")
else:
    print("Failed to train recognizer.")

3. Real-time Face Recognition

Class: RealTimeFaceRecognizer

• Attributes:

  • dataset_path: Path to the dataset folder.
  • trainer_path: Path to the trained recognizer file.
  • cascade_path: Path to the Haar Cascade XML file for face detection.
  • recognizer: LBPH Face Recognizer.
  • face_cascade: Cascade Classifier for face detection.
  • font: Font type for displaying text.
  • names: List of names extracted from the dataset.

• Methods:

  • get_names_from_dataset(): Retrieves names from the dataset.
  • load_recognizer(): Loads the trained recognizer.
  • detect_and_recognize_faces(img): Detects and recognizes faces in real-time.
  • start_recognition(): Initiates the real-time face recognition process.

Usage:

real_time_recognizer = RealTimeFaceRecognizer(
    dataset_path=args.dataset_path,
    trainer_path=args.trainer_path,
    cascade_path=args.cascade_path,
)

real_time_recognizer.start_recognition()

Conclusion

This face recognition system provides a comprehensive solution for capturing facial data, training a recognizer, and performing real-time face recognition. Users can customize parameters and paths based on their requirements.

---