Overview / Usage

About the Project: Attendance management system using face recognition

Accurate Attendance Recognition: CNNs are well-suited for image-based tasks, including face recognition. By training a CNN model with oneDNN, you can achieve accurate and reliable attendance recognition. The model can be trained on a large dataset of face images, enabling it to learn intricate facial features and patterns, resulting in robust attendance identification.

1. Real-time Processing: CNNs implemented with oneDNN leverage the performance optimizations and parallel processing capabilities of Intel processors. This enables real-time processing of attendance data, making it suitable for scenarios where attendance needs to be tracked and updated in real-time, such as in classrooms, workplaces, or events.
2. High Efficiency and Speed: oneDNN is designed to deliver high performance and efficiency in deep learning computations. It leverages Intel's advanced hardware optimizations, such as Intel Advanced Vector Extensions (AVX), to accelerate computations and improve throughput. This results in faster inference times and efficient attendance management.

There's are major question raise why we are using **OneDNN **for the following project ?

CNN stands for Convolutional Neural Network, which is a type of deep learning model commonly used for image recognition and computer vision tasks. OneDNN (formerly known as MKL-DNN) is a popular open-source library for deep learning, developed by Intel is base model used in this project. It provides highly optimized implementations of deep learning primitives, including convolutional layers, for various hardware architectures.

Using CNN models with OneDNN can offer several advantages and can be further utilised in enhancement of this project:

1. Performance Optimization: OneDNN is designed to leverage the underlying hardware architecture and optimize the computations involved in deep learning models. It can make efficient use of multi-core CPUs, vector instructions, and other hardware-specific features, resulting in improved performance and faster training or inference times.
2. Portability: OneDNN supports a wide range of hardware platforms, including CPUs, GPUs, and specialized accelerators. By using OneDNN, you can write your CNN models in a hardware-agnostic way, allowing them to run efficiently across different devices without significant code changes.
3. Integration with Existing Frameworks: OneDNN provides interfaces for popular deep learning frameworks such as TensorFlow and PyTorch, making it easy to integrate with your existing codebase. You can use the familiar APIs of these frameworks while benefiting from the optimised performance of OneDNN.
4. Flexibility: OneDNN offers a variety of pre-implemented functions and operations commonly used in CNN models, such as convolutions, pooling, and normalisation. It also supports different data formats and precision levels, allowing you to customize your models according to your specific requirements.

Overall, using CNN models with OneDNN can help you achieve faster and more efficient deep learning computations, regardless of the hardware platform you are targeting. It combines the flexibility of popular deep learning frameworks with the performance optimization provided by OneDNN's hardware-specific optimizations.

Methodology / Approach

Utilizing a CNN with oneDNN for attendance management brings accuracy, real-time processing, efficiency, scalability, and integration capabilities to your system. It enables you to develop a robust and high-performance attendance management solution that can handle the complexities of facial recognition and streamline attendance tracking processes.

Technologies Used

• OneDNN
• **numpy **
• sklearn.model_selection

Repository

https://github.com/Ragzoid/oneAPI-Intel-Face_Recognition_based_Attendance_management_system