Evolution of Hybrid Multi-modal Action Recognition: From DA-CNN+Bi-GRU to EfficientNet-CNN-ViT | Publicación