Article: A multimodal adaptive deep network for emotion recognition in artistic images Journal: International Journal of Reasoning-based Intelligent Systems (IJRIS) 2026 Vol.18 No.14 pp.70 - 88 Abstract: Recognising emotions in artworks is essential for digital galleries, personalised art recommendations, and art education. However, this task is challenging due to the abstract nature of images and subjective viewer interpretations, and existing methods often inadequately integrate visual content with textual descriptions. To address this issue, this paper proposes a multimodal adaptive emotion recognition network grounded in appraisal theory, featuring a gated adaptive fusion module that dynamically balances image and text contributions. An emotion-aware contrastive learning pre-training strategy is introduced to align cross-modal features. Experiments on the ArtEmis dataset show our method achieves 71.3% accuracy, surpassing state-of-the-art baselines by 2.4 percentage points. Ablation and case studies confirm the effectiveness and interpretability of each component. This work offers a promising solution for emotion understanding in art with demonstrable practical potential in controlled settings. Inderscience Publishers - linking academia, business and industry through research

Title: A multimodal adaptive deep network for emotion recognition in artistic images

Authors: Dangui Xu

Addresses: School of Art, Anhui University of Finance & Economics, Bengbu, 233030, China

Abstract: Recognising emotions in artworks is essential for digital galleries, personalised art recommendations, and art education. However, this task is challenging due to the abstract nature of images and subjective viewer interpretations, and existing methods often inadequately integrate visual content with textual descriptions. To address this issue, this paper proposes a multimodal adaptive emotion recognition network grounded in appraisal theory, featuring a gated adaptive fusion module that dynamically balances image and text contributions. An emotion-aware contrastive learning pre-training strategy is introduced to align cross-modal features. Experiments on the ArtEmis dataset show our method achieves 71.3% accuracy, surpassing state-of-the-art baselines by 2.4 percentage points. Ablation and case studies confirm the effectiveness and interpretability of each component. This work offers a promising solution for emotion understanding in art with demonstrable practical potential in controlled settings.

Keywords: artistic image emotion recognition; multimodal deep learning; adaptive fusion; appraisal theory; contrastive learning.

DOI: 10.1504/IJRIS.2026.154240

International Journal of Reasoning-based Intelligent Systems, 2026 Vol.18 No.14, pp.70 - 88

Received: 14 Mar 2026
Accepted: 10 Apr 2026
Published online: 17 Jun 2026 *

Title: A multimodal adaptive deep network for emotion recognition in artistic images

Keep up-to-date