Remote Heart Rate Estimation Based on Transformer with Multi-Skip Connection Decoder: Method and Evaluation in the Wild.

Heart rate is an essential vital sign to evaluate human health. Remote heart monitoring using cheaply available devices has become a necessity in the twenty-first century to prevent any unfortunate situation caused by the hectic pace of life. In this paper, we propose a new method based on the transformer architecture with a multi-skip connection biLSTM decoder to estimate heart rate remotely from videos. Our method is based on the skin color variation caused by the change in blood volume in its surface. The presented heart rate estimation framework consists of three main steps: (1) the segmentation of the facial region of interest (ROI) based on the landmarks obtained by 3DDFA; (2) the extraction of the spatial and global features; and (3) the estimation of the heart rate value from the obtained features based on the proposed method. This paper investigates which feature extractor performs better by captioning the change in skin color related to the heart rate as well as the optimal number of frames needed to achieve better accuracy. Experiments were conducted using two publicly available datasets (LGI-PPGI and Vision for Vitals) and our own in-the-wild dataset (12 videos collected by four drivers). The experiments showed that our approach achieved better results than the previously published methods, making it the new state of the art on these datasets.

Audio-Visual Speech and Gesture Recognition by Sensors of Mobile Devices.

Audio-visual speech recognition (AVSR) is one of the most promising solutions for reliable speech recognition, particularly when audio is corrupted by noise. Additional visual information can be used for both automatic lip-reading and gesture recognition. Hand gestures are a form of non-verbal communication and can be used as a very important part of modern human-computer interaction systems. Currently, audio and video modalities are easily accessible by sensors of mobile devices. However, there is no out-of-the-box solution for automatic audio-visual speech and gesture recognition. This study introduces two deep neural network-based model architectures: one for AVSR and one for gesture recognition. The main novelty regarding audio-visual speech recognition lies in fine-tuning strategies for both visual and acoustic features and in the proposed end-to-end model, which considers three modality fusion approaches: prediction-level, feature-level, and model-level. The main novelty in gesture recognition lies in a unique set of spatio-temporal features, including those that consider lip articulation information. As there are no available datasets for the combined task, we evaluated our methods on two different large-scale corpora-LRW and AUTSL-and outperformed existing methods on both audio-visual speech recognition and gesture recognition tasks. We achieved AVSR accuracy for the LRW dataset equal to 98.76% and gesture recognition rate for the AUTSL dataset equal to 98.56%. The results obtained demonstrate not only the high performance of the proposed methodology, but also the fundamental possibility of recognizing audio-visual speech and gestures by sensors of mobile devices.