A comparative study on facial expression recognition using MobileNetV2, VGG-16, ResNet and Swin Transformer
Bibliographic record
Abstract
In the modern era, the implementation of computer vision technologies has become increasingly prevalent across various domains. In many fields, such as security and identification, computer vision technologies are widely implemented and have gradually become increasingly prevalent topics. Recent advancements in facial emotion recognition models have proved to be significant for the development of different deep learning models in past research. Undoubtedly, Convolutional Neural Networks (CNNs) have long been the dominant approach for such facial recognition tasks, especially for emotion classification tasks. Instead of the traditional CNNs, some other researchers prefer using vit such as (Mobile Swim Windows Attention) to solve such problems because they argue that it could extract more global features rather than local features extracted by CNN. To determine the advantages and disadvantages on these models for the performance tasks, this paper presents a comparative analysis of CNN-based models - MobileNetV2, VGG16, ResNet - against the Swin Transformer, a Vision Transformer (ViT) model. Using the FER2013 datasets with 48x48 pixels images that are all classified into 7 different emotion categories and implementing the augmentation techniques to improve the quality of the datasets, we utilized the pretrained models, train them with at least 20 epochs, and evaluated the performance of classical convolutional neural networks and Vision Transformer on this dataset based on accuracy, F1 Score, Precision, and Recall with accuracy, loss, and confusion matrix visualization.
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".