In audiovisual emotion recognition, representational learning is a research direction receiving considerable attention, and the key lies in constructing effective affective representations with both consistency and variability. However, there are still many challenges to accurately realize affective representations. For this reason, in this paper we proposed a cross-modal audiovisual recognition model based on a multi-head cross-attention mechanism. The model achieved fused feature and modality alignment through a multi-head cross-attention architecture, and adopted a segmented training strategy to cope with the modality missing problem. In addition, a unimodal auxiliary loss task was designed and shared parameters were used in order to preserve the independent information of each modality. Ultimately, the model achieved macro and micro F1 scores of 84.5% and 88.2%, respectively, on the crowdsourced annotated multimodal emotion dataset of actor performances (CREMA-D). The model in this paper can effectively capture intra- and inter-modal feature representations of audio and video modalities, and successfully solves the unity problem of the unimodal and multimodal emotion recognition frameworks, which provides a brand-new solution to the audiovisual emotion recognition.
Emotions ; Humans ; Attention ; Algorithms

Objective:To explore the accuracy of computer optometric diopter before mydriasis in predicting artificial optometric diopter after mydriasis in the current ophthalmic standard of civil aviation recruitment, to evaluate the rationality of "0.75 D or more beyond the standard" in the current recruitment standard, and to provide a scientific basis for adjusting the recruitment diopter standard.Methods:Cluster sampling was used to select students recruited by Civil Aviation Flight University of China in Sichuan area from 2017 to 2021, including high school students and college student candidates. The values of computer optometric diopter before mydriasis and optometric diopter after mydriasis were collected, and the linear regression model was fitted to predict the optometric diopter value after mydriasis with computer optometric diopter before mydriasis as independent variable.Results:A total of 2 567 recruited students, all male, with an average age of (18.20±1.47) years were enrolled in the physical examination. There was a high correlation between computer optometric diopter before mydriasis and optometric diopter after mydriasis ( rS=0.856, P<0.001). After linear regression fitting and calculation of the predicted range of post-dilated optometric diopter, it was found that pre-dilated computer optometric diopter could explain 80.5% of the variation degree of post-dilated optometric diopter. When the computer optometric diopter before mydriasis was in the range of [-7.000,+1.500] D, the prediction accuracy of this model was ≥83.3%, and the fitting effect was good. When the computer optometric diopter before mydriasis was higher than -6.125 D, the prediction interval of 95% was higher than -4.500 D, which was beyond the recruitment standard. Conclusions:For myopic students, the predicted value of post-dilated optometric diopter exceeds the current recruitment standard if the computer optometric diopter before mydriasis exceeds -6.125 D. It is suggested to amend the corresponding item in the standard to "if the diopter exceeds 1.750 D or more, the subject can be terminated other examinations".

Objective:To explore the accuracy of computer optometric diopter before mydriasis in predicting artificial optometric diopter after mydriasis in the current ophthalmic standard of civil aviation recruitment, to evaluate the rationality of "0.75 D or more beyond the standard" in the current recruitment standard, and to provide a scientific basis for adjusting the recruitment diopter standard.Methods:Cluster sampling was used to select students recruited by Civil Aviation Flight University of China in Sichuan area from 2017 to 2021, including high school students and college student candidates. The values of computer optometric diopter before mydriasis and optometric diopter after mydriasis were collected, and the linear regression model was fitted to predict the optometric diopter value after mydriasis with computer optometric diopter before mydriasis as independent variable.Results:A total of 2 567 recruited students, all male, with an average age of (18.20±1.47) years were enrolled in the physical examination. There was a high correlation between computer optometric diopter before mydriasis and optometric diopter after mydriasis ( rS=0.856, P<0.001). After linear regression fitting and calculation of the predicted range of post-dilated optometric diopter, it was found that pre-dilated computer optometric diopter could explain 80.5% of the variation degree of post-dilated optometric diopter. When the computer optometric diopter before mydriasis was in the range of [-7.000,+1.500] D, the prediction accuracy of this model was ≥83.3%, and the fitting effect was good. When the computer optometric diopter before mydriasis was higher than -6.125 D, the prediction interval of 95% was higher than -4.500 D, which was beyond the recruitment standard. Conclusions:For myopic students, the predicted value of post-dilated optometric diopter exceeds the current recruitment standard if the computer optometric diopter before mydriasis exceeds -6.125 D. It is suggested to amend the corresponding item in the standard to "if the diopter exceeds 1.750 D or more, the subject can be terminated other examinations".