Xingxu Yao

TL;DR: An Attention-aware Polarity Sensitive Embedding (APSE) network is introduced to learn affective representations in an end-to-end manner and a weighted emotion-pair loss is presented to take the inter- and intra-polarity relationships of the emotional labels into consideration.

TL;DR: Recently, with the rapid advancement of emotional intelligence and the explosive growth of visual data, extensive research efforts have been dedicated to affective image content analysis (AICA) as mentioned in this paper.

TL;DR: An adaptive sentiment similarity loss is designed, which is able to embed affective images considering the emotion polarity and adaptively adjust the margin between different image pairs, and a unified multi-task deep framework is developed to simultaneously optimize both retrieval and classification goals.

TL;DR: End-to-end matching between image and music based on emotions in the continuous valence-arousal (VA) space is studied, demonstrating the superiority of CDCML for emotion-based image andMusic matching as compared to the state-of-the-art approaches.

TL;DR: Zhang et al. as discussed by the authors proposed cross-modal deep continuous metric learning (CDCML) to learn a shared latent embedding space which preserves the crossmodal similarity relationship in the continuous matching space.