Complementary Global–Local Feature Fusion and Ensemble Refinement for Facial-Expression Recognition on FER2013

Дата публикации: 06-10-2026 13:45:02

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 captures global facial semantics, while EfficientNetB3 extracts complementary local texture features. Their fused representation is used for seven-class facial-expression classification. The classification head is first trained with targeted feature-space SMOTE, and the EfficientNetB3 branch is then partially fine-tuned. Five-view test-time augmentation (TTA) is further incorporated at inference, together with an independently trained ConvNeXt-Tiny branch to provide additional architectural diversity. Ensemble weights are selected using a held-out validation set, while the test partition is reserved for final evaluation. On the 7178-image FER2013 test partition, the resulting ensemble achieves 76.51% accuracy, 75.77% macro F1, 76.32% weighted F1, 0.7158 Cohen’s kappa, and 0.9574 macro ROC AUC. The results demonstrate the potential of combining complementary global–local representations, inference augmentation, and ensemble refinement for facial-expression recognition on FER2013.

Схожие новости

#	Наименование новости	Тональность	Информативность	Дата публикации
1	Revolutionizing Alzheimer’s Diagnosis: A Hybrid Deep Learning Approach for Enhanced MRI Analysis	0	8	06-10-2026
2	NVIDIA’s Groq 3 LPU Accelerators for Heterogeneous AI Compute at Hot Chips 2026	0	8.78	25-08-2026
3	https://get.faceapp.com	0	10	03-10-2026
4	Google’s TPUv8s for Training and Inference at Hot Chips 2026	0	9.98	26-08-2026
5	Microsoft want to record your face and feed it to a generative AI, telling it to go off and design a game from your frown	0	4.58	23-09-2026
6	An Industry-Proof Over-the-Air Vehicle-in-the-Loop Approach for the Virtual Verification and Validation of Automotive Radar	0	31.43	11-09-2026
7	Beyond Static Coexistence: Architecting Non-Stationary RFFE Systems for Multi-State Devices	0	30	12-08-2026
8	Generative AI 3D building from Autocad PDFs:	0	9.18	02-10-2026
9	Daily Hacker News for 2026-09-29	0	11.19	30-09-2026

Классификация: Пресс-релизы. Схожих патентов: 0. Схожих новостей: 9. Тональность: 0. Информативность: 18.56. Источник: engagedscholarship.csuohio.edu.