Emotion Recognition from Video Frame Sequence using Face Mesh and Pre-Trained Models of Convolutional Neural Network
Bibliographic record
Abstract
Emotions are a collection of subjective cognitive experiences and psychological and physiological characteristics that express a wide range of feelings, thoughts, and behaviors in human interaction. Emotions can be represented through several means, such as facial expressions, tone of voice, and behavior. Deep Learning (DL) research has focused on incorporating facial expressions. Images with facial expressions are commonly used as data input for the DL model. Unfortunately, most DL models in Facial Emotion Recognition (FER) use static images. This method does not take into consideration all conceivable facial expressions. The static image of facial expressions is insufficient for recognizing emotions, but a sequential image from a video is required. In this study, we extract MediaPipe’s face mesh feature, the state-of-the-art multidimensional expression key points embedded in the video image sequence. Furthermore, we feed sequence image data into the pre-trained Convolutional Neural Network (CNN) model. The data we used is from The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) with the emotion classes of “Anger,” “Fearful,” “Happy,” and “Sad.” For this specific FER task, we found that the best pre-trained CNN model achieved 92.8% accuracy (using the VGG-19 model), with the fastest runtime of $\sim2.3$ seconds (achieved using the SqueezeNet model).
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 | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| 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".