In this paper, we propose an integrated approach to robot vision: a key frame-based skeleton feature estimation and action recognition network (KFSENet) that incorporates action recognition with face and emotion recognition to enable social robots to engage in more personal interactions. Instead of extracting the human skeleton features from the entire video, we propose a key frame-based approach for their extraction using pose estimation models. We select the key frames using the gradient of a proposed total motion metric that is computed using dense optical flow. We use the extracted human skeleton features from the selected key frames to train a deep neural network (i.e., the double-feature double-motion network (DDNet)) for action recognition. The proposed KFSENet utilizes a simpler model to learn and differentiate between the different action classes, is computationally simpler and yields better action recognition performance when compared with existing methods. The use of key frames allows the proposed method to eliminate unnecessary and redundant information, which improves its classification accuracy and decreases its computational cost. The proposed method is tested on both publicly available standard benchmark datasets and self-collected datasets. The performance of the proposed method is compared to existing state-of-the-art methods. Our results indicate that the proposed method yields better performance compared with existing methods. Moreover, our proposed framework integrates face and emotion recognition to enable social robots to engage in more personal interaction with humans.

This letter presents a novel error correction module using a Bag-of-Words model and deep neural networks to improve the accuracy of cloud-based speech-to-text services on recognition tasks of non-native speakers with foreign accents. The Bag-of-Words model transforms text into input vectors for the deep neural network, which is trained using typical sentences in the curriculum for elementary schools in Vietnam and the Google Speech-to-Text data for those sentences. The trained network is then used for real-time error correction on a humanoid robot and yields 18% better accuracy than Google Speech-to-Text.