Face Recognition in Video Streams and its Application in Freedom Fighters Discovery - A Machine Learning Approach
Vani Perumal · 2020
Face Recognition is an automated process implemented by using various tools and techniques, which is capable of detecting and identifying human faces from any image or video stream. Machine learning is an approach, which involves in training a machine for prediction and decision making. It simulates the function of a brain through proper training. In this paper, a machine learning approach is used to predict the faces of Indian freedom fighters from any video stream given as an input. To accomplish this task, the system must be trained before with an appropriate and adequate numbers of facial dataset. The training dataset is also created from a video stream, which is named as training video. Provided, that particular video stream must contain only one specific freedom fighter's images of different stages and different styles. From the training video stream, image frames consist of their face were identified, extracted and stored for each person. Then these frames of faces were used to train the machine. After proper training, a model was created and this model was used for prediction. In this work, Indian freedom fighters' images were used for training and prediction. The same approach could be utilized for the prediction of any person or list of persons from a video stream after procuring proper training.