Improving Video Vision Transformer for Deepfake Video Detection Using Facial Landmark, Depthwise Separable Convolution and Self Attention | IEEE Journals & Magazine | IEEE Xplore