This project focuses on developing a comprehensive face authentication system that integrates face verification and anti-spoofing capabilities. A diverse dataset was prepared, featuring individuals from various demographics and capturing a wide range of face shapes. This dataset was enhanced using PyTorch through image augmentation techniques, such as contrast adjustments, to improve variability and robustness. Care was taken to ensure that the dataset was balanced and that all images maintained consistent dimensions, which is crucial for effective model training.

The MobileNetV2 architecture may be selected for its efficiency and accuracy. Fine-tuning should be conducted by increasing the number of epochs and adjusting the learning rate, allowing the model to learn effectively from the diverse dataset. Different lighting conditions may be also considered to ensure robust performance in real-world scenarios.

The expected outcome of this project is a reliable face authentication system capable of identifying spoofed faces across various devices, independent of pixel resolution. By training the model on a well-curated dataset with applied augmentations, it aims to effectively detect unauthorized access attempts while maintaining high accuracy in diverse environments. This capability is essential for enhancing security in applications requiring face verification, ultimately leading to a more secure user experience. The combination of thorough dataset preparation, model optimization, and robust testing across various conditions will ensure effective performance in real-world applications.

1. Clone the Repository: Start by cloning the project’s repository.
2. Create a Branch: Use Git to create your own branch for the task.
3. Generate a Pull Request: After completing your coding, submit your task by generating a pull request.

https://openforge.gov.in/plugins/git/ddc/face_detection.git

Organization Type : Government Publisher Name : DigiLocker, National e-Governance Division (NeGD)