Gaze estimation methods encounter significant performance deterioration when being evaluated across different domains, because of the domain gap between the testing and training data. Existing methods try to solve this issue by reducing the deviation of data distribution, however, they ignore the existence of label deviation in the data due to the acquisition mechanism of the gaze label and the individual physiological differences. In this paper, we first point out that the influence brought by the label deviation cannot be ignored, and propose a gaze label alignment algorithm (GLA) to eliminate the label distribution deviation. Specifically, we first train the feature extractor on all domains to get domain invariant features, and then select an anchor domain to train the gaze regressor. We predict the gaze label on remaining domains and use a mapping function to align the labels. Finally, these aligned labels can be used to train gaze estimation models. Therefore, our method can be combined with any existing method. Experimental results show that our GLA method can effectively alleviate the label distribution shift, and SOTA gaze estimation methods can be further improved obviously.

Since the collection of labeled gaze data is laborious and time-consuming, methods which can learn generalizable features by leveraging large-scale available unlabeled data are desirable. In recent years, we have witnessed the tremendous capabilities of diffusion models in generating images as well as their potential in feature representation learning. In this paper, we investigate whether they can acquire discriminative representations for gaze estimation via generative pre-training. To achieve this goal, we propose a self-supervised learning framework with diffusion models for gaze estimation, called GazeDiff. Specifically, we utilize a conditional diffusion model to generate target image with gaze direction specified by the reference image as the pre-training task. To facilitate the diffusion model to learn gaze related features as condition, we propose a disentangling feature learning strategy, which first learns appearance feature, head pose feature, and eye direction feature respectively, and then combines them as the conditional features. Extensive experiments demonstrate denoising diffusion models are also good general gaze feature learners.

Gaze estimation methods often experience significant performance degradation when evaluated across different domains, due to the domain gap between the testing and training data. Existing methods try to address this issue using various domain generalization approaches, but with little success because of the limited diversity of gaze datasets, such as appearance, wearable, and image quality. To overcome these limitations, we propose a novel framework called CLIP-Gaze that utilizes a pre-trained vision-language model to leverage its transferable knowledge. Our framework is the first to leverage the vision-and-language cross-modality approach for gaze estimation task. Specifically, we extract gaze-relevant feature by pushing it away from gaze-irrelevant features which can be flexibly constructed via language descriptions. To learn more suitable prompts, we propose a personalized context optimization method for text prompt tuning. Furthermore, we utilize the relationship among gaze samples to refine the distribution of gaze-relevant features, thereby improving the generalization capability of the gaze estimation model. Extensive experiments demonstrate the excellent performance of CLIP-Gaze over existing methods on four cross-domain evaluations.

In this article, a new dataset, the driver emotion facial expression (DEFE) dataset for drivers’ spontaneous emotions analysis is introduced. The dataset includes facial expression recordings from 60 participants during driving. After watching a selected video-audio clip to elicit a specific emotion, each participant completed the driving tasks in the same driving scenario and rated his/her emotional responses during the driving processes from the aspects of dimensional emotion method and discrete emotion method. The study also conducted classification experiments to recognize the scales of arousal, valence, dominance, as well as the emotion category and intensity to establish baseline results for the proposed dataset. Furthermore, this paper compared emotion recognition results difference through facial expressions between dynamic driving and static life scenarios. The results showed that dynamic driving and static life datasets were different in emotion recognition results. To further explore the reasons for the difference in emotion recognition results, the analysis from the AU (action unit) presence perspective was studied. The results showed significant differences in the AUs presence of facial expressions between dynamic driving and static life scenarios, indicating that drivers’ facial expressions may be affected by the driving task to influence the recognition of drivers’ emotions through facial expressions. Therefore, to accurately recognize the drivers’ emotions to establish a reliable emotion-aware human-machine interaction system, thereby improving driving safety and comfort, publishing a human emotion dataset specifically for the driver is necessary. The proposed dataset will be publicly available so that researchers worldwide can use it to develop and examine their driver emotion analysis methods. To the best of our knowledge, this is currently the only public driver facial expression dataset.

Driver’s emotion recognition is vital to improving driving safety, comfort, and acceptance of intelligent vehicles. This article presents a cognitive-feature-augmented driver emotion detection method that is based on emotional cognitive process theory and deep networks. Different from the traditional methods, both the driver’s facial expression and cognitive process characteristics (age, gender, and driving age) were used as the inputs of the proposed model. Convolutional techniques were adopted to construct the model for driver’s emotion detection simultaneously considering the driver’s facial expression and cognitive process characteristics. A driver’s emotion data collection was carried out to validate the performance of the proposed method. The collected dataset consists of 40 drivers’ frontal facial videos, their cognitive process characteristics, and self-reported assessments of driver emotions. Another two deep networks were also used to compare recognition performance. The results prove that the proposed method can achieve well detection results for different databases on the discrete emotion model and dimensional emotion model, respectively.