Fer2013 Recognition - ResNet18 With Tricks
Xiaojian Yuan
Abstract
This work is the final project of the Computer Vision Course of USTC. However, I achieve the highest single-network classification accuracy on FER2013 based on ResNet18. To my best knowledge, this work achieves state-of-the-art single-network accuracy of 73.70 % on FER2013 without using extra training data, which exceeds the previous work [1] of 73.28%.
Related Papers
Multimodal Prompt Alignment for Facial Expression Recognition2025-06-26EfficientFER: EfficientNetv2 Based Deep Learning Approach for Facial Expression Recognition2025-06-02TKFNet: Learning Texture Key Factor Driven Feature for Facial Expression Recognition2025-05-15Achieving 3D Attention via Triplet Squeeze and Excitation Block2025-05-09SDAFE: A Dual-filter Stable Diffusion Data Augmentation Method for Facial Expression Recognition2025-04-06Evaluating Facial Expression Recognition Datasets for Deep Learning: A Benchmark Study with Novel Similarity Metrics2025-03-26Disentangled Source-Free Personalization for Facial Expression Recognition with Neutral Target Data2025-03-26V-NAW: Video-based Noise-aware Adaptive Weighting for Facial Expression Recognition2025-03-20