Video-based face recognition and image synthesis from rotating head frames using nonlinear manifold learning by neural networks

Real‐time face recognition based on pre‐identification and multi‐scale classification

“…Overall, advanced research studies concerning face recognition are mostly based on machine learning, which finds the objects by traversing the whole image with a search box. Thus, large objects are more likely to be searched because there are less interfering pixels in the image [3].…”

“…Recently, deep learning‐based algorithms have been widely applied in the computer vision research area [3]. Wang et al [12] proposed a joint method of priori convolutional neural networks at the superpixel level (called as ‘priori s‐CNNs’) and soft restricted context transfer.…”

“…Traditional face recognition methods recognise faces by detecting human faces, tracking the faces, capturing a frame of the image, and then extracting features and matching faces in static images. For instance, Hamedani et al [3] proposed a video‐based face recognition method that first detected and tracked faces in videos and then captured video frames of a subject rotating his/her head. The manifolds of video frames embedded in a high‐dimensional video space were extracted using neural network‐based models.…”

“…The reason is that the basic principle of machine learning for detecting an object is to traverse the whole picture with a search box to find the most similar regions to the object. Large objects are more likely to be searched because there are less interfering pixels in the image [3]. Therefore, instead of searching faces directly, we obtain the area of the moving pedestrian by aggregate channel feature (ACF) features and support vector machine (SVM) classifier and then estimate the face area based on the proportion of human faces in the pedestrian area to reduce the search range of the detector.…”