Aug 2026· 2026 7th International Conference on Computer Vision and Data Mining (ICCVDM)· pp. 187-190· 0 citations· 7 references
Abstract
Facial expression recognition (FER) remains challenging due to occlusion, pose variations, and complex background interference. To address these issues, this paper proposes a novel FER framework that integrates global collaborative features with local enhancement features. The model adopts ConvNeXt as the backbone and introduces a Global Feature Collaboration Branch (GFCB) and a Local Feature Enhancement Branch (LFEB). The global branch captures channel-wise and spatial-wise dependencies through a Channel Interaction Module and a Spatial Interaction Module, enabling comprehensive contextual modeling. The local branch employs asymmetric convolution and attention mechanisms to extract fine-grained facial texture features and improve robustness to local distortions. A feature fusion module is designed to integrate global semantics and local details adaptively. Extensive experiments on RAF-DB and AffectNet datasets demonstrate that the proposed method achieves 90.52% and 62.89% accuracy, respectively, outperforming state-of-the-art methods and verifying its effectiveness and robustness.
A hybrid model which uses several ResNet50 mixed with VGG16 backbones to improve the local facial features with the global context for a robust FER solu-tion, and shows superior robustness to occlusions and pose, even under cross-dataset testing.
A. Ouza, Mohamed El Ghmary, A. Choukri· Multimedia tools and applica...· 0 citations
Facial-expression recognition (FER) on FER2013 remains challenging because of low-resolution images, class imbalance, and label ambiguity. This study presents a global–local feature-fusion framework that integrates complementary representations with validation-based ensemble refinement. A frozen DINOv2 ViT-Base capture...
H. Shahzad, Hassan A. Ahmed· Information· 0 citations
In recent years, Facial Expression Recognition (FER) has gained attention as a crucial area of computer vision. However, FER methods depend on single-source facial data that often experience reduced accuracy and instability because of challenges such as facial occlusion and variations in illumination. Moreover, existin...