Face images contain many important biological characteristics. The research directions of face images mainly include face age estimation, gender judgment, and facial expression recognition. Taking face age estimation as an example, the estimation of face age images through algorithms can be widely used in the fields of biometrics, intelligent monitoring, human-computer interaction, and personalized services. With the rapid development of computer technology, the processing speed of electronic devices has greatly increased, and the storage capacity has been greatly increased, allowing deep learning to dominate the field of artificial intelligence. Traditional age estimation methods first design features manually, then extract features, and perform age estimation. Convolutional neural networks (CNN) in deep learning have incomparable advantages in processing image features. Practice has proven that the accuracy of using convolutional neural networks to estimate the age of face images is far superior to traditional methods. However, as neural networks are designed to be deeper, and networks are becoming larger and more complex, this makes it difficult to deploy models on mobile terminals. Based on a lightweight convolutional neural network, an improved ShuffleNetV2 network based on the mixed attention mechanism (MA-SFV2: Mixed Attention-ShuffleNetV2) is proposed in this paper by transforming the output layer, merging classification and regression age estimation methods, and highlighting important features by preprocessing images and data augmentation methods. The influence of noise vectors such as the environmental information unrelated to faces in the image is reduced, so that the final age estimation accuracy can be comparable to the state-of-the-art.
Remote photoplethysmography (rPPG) is a video-based non-contact heart rate measurement technology. It is a fact that most existing rPPG methods fail to deal with the spatiotemporal features of the video, which is significant for the extraction of the rPPG signal. In this paper, we propose a 3D central difference convolutional network (CDCA-rPPGNet) to measure heart rate, with an attention mechanism to combine spatial and temporal features. First, we crop and stitch the region of interest together through facial landmarks. Next, the high-quality regions of interest are fed to CDCA-rPPGNet based on a central difference convolution, which can enhance the spatiotemporal representation and capture rich relevant time contexts by collecting time difference information. In addition, we integrate the attention module into the neural network, aiming to strengthen the ability of the neural network to extract video channels and spatial features, so as to obtain more accurate rPPG signals. In summary, the three main contributions of this paper are as follows: (1) the proposed network base on central difference convolution could better capture the subtle color changes to recover the rPPG signals; (2) the proposed ROI extraction method provides high-quality input to the network; (3) the attention module is used to strengthen the ability of the network to extract features. Extensive experiments are conducted on two public datasets—the PURE dataset and the UBFC-rPPG dataset. In terms of the experiment results, our proposed method achieves 0.46 MAE (bpm), 0.90 RMSE (bpm) and 0.99 R value of Pearson’s correlation coefficient on the PURE dataset, and 0.60 MAE (bpm), 1.38 RMSE (bpm) and 0.99 R value of Pearson’s correlation coefficient on the UBFC dataset, which proves the effectiveness of our proposed approach.
Micro-Doppler-based continuous human motion recognition (HMR) has gained considerable attention recently. However, existing methods mainly rely on individual recurrent neural network or sliding-window-based approaches, which makes them hard to effectively exploit all the temporal information to predict motions. Additionally, they need to represent the raw radar data into other domains and then perform feature extraction and classification. Thus, the representation cannot be optimized, and its high computational complexity and independence from learning model make the network consume significant time. In this article, to address these issues, we propose a new end-to-end network that uses radar radios to recognize continuous motion. Specifically, the fusion layer fuses the raw I & Q radar data without the need of representations, and it is integrated with subsequent networks in an end-to-end manner for jointly optimization. Moreover, the attention-based encoder-decoder structure encodes the fused data and selects useful temporal information for recognition, which guarantees the effective use of all the temporal information. The experiments show that in continuous HMR, the proposed network outperforms existing methods in terms of accuracy and inference time.
The application of micro-Doppler-based continuous human motion recognition (HMR) is greatly hindered by non-target micro motion interference, due to the deformation of micro-Doppler signatures of target human motion caused by such interference. In this paper, we propose a novel continuous HMR method using micro-Doppler signatures that can work in the scenario with non-target micro motion interference. Specifically, a signal preprocessing architecture is designed, where the empirical mode decomposition is employed to remove the interference in radar raw signal and the multiwindow time-frequency representation is used to generate the time-frequency distribution (TFD) with high concentration. Moreover, a tailored network, that integrates multiscale squeeze-and-excitation network for feature sequence extraction, stacked bidirectional long short-term memory for sequence labeling and connectionist temporal classification algorithm for label transcription, is employed to recognize continuous human motion from TFD. The experimental results show that the proposed method outperforms existing methods in terms of recognition accuracy and generalization.
Human motion classification based on micro-Doppler effect has been widely used in various fields. However, the motion classification performance would be greatly degraded if the wireless environment has non-target micro-motion interference. In this case, the interference signal aliases with the signal of target human motions and then generates cross-terms, making the signals hard to be used to identify target human motions. Existing methods do not consider this non-target micro-motion interference and have poor resistance to such interference. In this paper, we propose a target human motion classification system that can work in the scenarios with non-target micro-motion interference. Specifically, we build a continuous wave radar transceiver working in a low-frequency radar band using the software defined radio equipment Universal Software Radio Peripheral (USRP) N210 to collect signals. Moreover, we use Empirical Mode Decomposition and S-transform successively to remove non-target micro-motion interference and improve the time-frequency resolution of the raw signal. Then, an Energy Aggregation method based on S-method is proposed, which can suppress cross-terms and background noise. Furthermore, we extract a set of features and classify four human motions by adopting Bagged Trees. Extensive experiments using the test-bed show that under the scenarios with non-target micro-motion interference, 97.3% classification accuracy can be achieved.
Studies have shown that ordinary color cameras can detect the subtle color changes of the skin caused by the heartbeat cycle. Therefore, cameras can be used to remotely monitor the pulse in a non-contact manner. The technology for non-contact physiological measurement in this way is called remote photoplethysmography (rPPG). Heart rate variability (HRV) analysis, as a very important physiological feature, requires us to be able to accurately recover the peak time locations of the rPPG signal. This paper proposes an efficient spatiotemporal attention network (ESA-rPPGNet) to recover high-quality rPPG signal for heart rate variability analysis. First, 3D depth-wise separable convolution and a structure based on mobilenet v3 are used to greatly reduce the time complexity of the network. Next, a lightweight attention block called 3D shuffle attention (3D-SA), which integrates spatial attention and channel attention, is designed to enable the network to effectively capture inter-channel dependencies and pixel-level dependencies. Moreover, ConvGRU is introduced to further improve the network’s ability to learn long-term spatiotemporal feature information. Compared with existing methods, the experimental results show that the method proposed in this paper has better performance and robustness on the remote HRV analysis.