🤖 AI Summary
Existing facial expression recognition (FER) datasets predominantly provide only discrete emotion category labels, failing to capture fine-grained, continuous affective variations. Although some works incorporate Valence-Arousal (VA) annotations, the Dominance (D) dimension remains largely absent. This paper addresses this gap by introducing the first complete, human-verified three-dimensional Valence-Arousal-Dominance (VAD) continuous annotation for the FER2013 dataset. We further propose an orthogonal convolution-based ResNet regression architecture that enforces feature orthogonality to improve VAD value prediction accuracy. Experiments demonstrate that, despite its high annotation difficulty, the D dimension is effectively learnable; orthogonal convolutions significantly enhance predictive performance across all three dimensions—particularly for Dominance. The released VAD-annotated FER2013 dataset and open-source code establish a new benchmark for multidimensional affective computing, enabling more precise, granular emotion analysis in applications such as intelligent education.
📝 Abstract
Current FER (Facial Expression Recognition) dataset is mostly labeled by emotion categories, such as happy, angry, sad, fear, disgust, surprise, and neutral which are limited in expressiveness. However, future affective computing requires more comprehensive and precise emotion metrics which could be measured by VAD(Valence-Arousal-Dominance) multidimension parameters. To address this, AffectNet has tried to add VA (Valence and Arousal) information, but still lacks D(Dominance). Thus, the research introduces VAD annotation on FER2013 dataset, takes the initiative to label D(Dominance) dimension. Then, to further improve network capacity, it enforces orthogonalized convolution on it, which extracts more diverse and expressive features and will finally increase the prediction accuracy. Experiment results show that D dimension could be measured but is difficult to obtain compared with V and A dimension no matter in manual annotation or regression network prediction. Secondly, the ablation test by introducing orthogonal convolution verifies that better VAD prediction could be obtained in the configuration of orthogonal convolution. Therefore, the research provides an initiative labelling for D dimension on FER dataset, and proposes a better prediction network for VAD prediction through orthogonal convolution. The newly built VAD annotated FER2013 dataset could act as a benchmark to measure VAD multidimensional emotions, while the orthogonalized regression network based on ResNet could act as the facial expression recognition baseline for VAD emotion prediction. The newly labeled dataset and implementation code is publicly available on https://github.com/YeeHoran/VAD-Net .