Cloud based architecture for Face Recognition in Django with Machine Learning
Shubhankar Bhardwaj, Saurabh Sharma, Vinay Pratap Singh · 2023
If we compare a face identification system to others like fingerprint, iris scanning, signature, etc., its applicability is simpler and its operating range is bigger than others. The face has a more pronounced structure and a larger surface area than other biometrics, and it can be collected without any physical touch.. We suggested combining Django and machine learning to create a cloud-based architecture for face recognition. The system recognises several faces in live captured images using a mix of techniques, including face identification with a histogram of oriented gradients (HOG), image augmentation, extraction of facial embedding, and classification of images using linear SVM. The local system camera is used to identify faces on individuals. Face detection, Face landmark, Face recognition is done first. The system is then updated to include a face classification technique that makes use of linear SVM. A database that we created with the help of 10 people is used to test the system. Within the desired parameters, the tested system performs acceptable for facial recognition. On the provided data, the linear SVM algorithm's accuracy is 99%. The system can also identify several faces in live photos that have been captured.