A channel attention and feature manipulation network for facial expression recognition
Bibliographic record
Abstract
Facial expression conveys a variety of emotional and intentional message from human beings, and automated facial expression recognition (FER) has become an ongoing and promising research topic in the field of computer vision. However, the primary challenge of FER is learning to discriminate similar features among different emotion categories. In this paper, a hybrid architecture using Efficient Channel Attention (ECA) residual network ResNet-18, and feature manipulation network is proposed to tackle the above challenge. First, the ECA residual network effectively extract input features with local cross-channel interaction. Then, the feature decomposition network (FDN), feature reconstruction network (FRN) modules are added to decompose and aggregate latent features for enhancing the compactness of intra-category features and discrimination of inter-category features. Finally, an expression prediction network is connected to FRN to draw the final expression classification result. To examine the efficacy of the suggested approach, the model is trained independently using in-the-lab (CK+) and in-the-wild (RAF-DB) datasets. Several important evaluation metrics such as confusion matrix, Grad-CAM are reported, and the ablation study is conducted for demonstrating the efficacy and interpretability of the proposed network. It achieves the state-of-the-art accuracy compared to the existing facial recognition work, at 99.70% in CK+ and 89.17% in RAF-DB.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".